diff --git a/.claude-plugin/marketplace.json b/.claude-plugin/marketplace.json
new file mode 100644
index 0000000..d537c8a
--- /dev/null
+++ b/.claude-plugin/marketplace.json
@@ -0,0 +1,26 @@
+{
+  "name": "kagi-cli",
+  "owner": {
+    "name": "Microck",
+    "url": "https://github.com/Microck"
+  },
+  "metadata": {
+    "description": "Marketplace for the Kagi CLI plugin.",
+    "version": "1.0.0"
+  },
+  "plugins": [
+    {
+      "name": "kagi",
+      "source": "./",
+      "description": "Kagi CLI skills for research, page content, Assistant, monitoring, and account configuration.",
+      "version": "0.15.0",
+      "author": {
+        "name": "Microck",
+        "url": "https://github.com/Microck"
+      },
+      "homepage": "https://github.com/Microck/kagi-cli#readme",
+      "license": "MIT",
+      "keywords": ["kagi", "search", "cli", "research", "assistant"]
+    }
+  ]
+}
diff --git a/.claude-plugin/plugin.json b/.claude-plugin/plugin.json
new file mode 100644
index 0000000..109e9a2
--- /dev/null
+++ b/.claude-plugin/plugin.json
@@ -0,0 +1,14 @@
+{
+	"$schema": "https://json.schemastore.org/claude-code-plugin-manifest.json",
+  "name": "kagi",
+  "version": "0.15.0",
+  "description": "Kagi CLI skills for research, page content, Assistant, monitoring, and account configuration.",
+  "author": {
+    "name": "Microck",
+    "url": "https://github.com/Microck"
+  },
+  "homepage": "https://github.com/Microck/kagi-cli#readme",
+  "repository": "https://github.com/Microck/kagi-cli",
+  "license": "MIT",
+  "keywords": ["kagi", "search", "cli", "research", "assistant"]
+}
diff --git a/CHANGELOG.md b/CHANGELOG.md
index 67458c7..513fef5 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -11,7 +11,8 @@ Before `1.0.0`, breaking changes may still ship in minor releases.
 
 ### Added
 
-- `kagi skills` now includes focused guides for research, page content, Assistant conversations, monitoring and automation, and account configuration.
+- Install the repository as a Claude Code plugin through its bundled marketplace manifest.
+- `kagi skills` now includes a unified `kagi-usage` guide for search and page content, a `kagi-ai` guide for spending Assistant allowance, and focused guides for Assistant conversations, monitoring and automation, and account configuration.
 
 ### Fixed
 
diff --git a/README.md b/README.md
index d0ed7a1..65e316a 100644
--- a/README.md
+++ b/README.md
@@ -170,7 +170,7 @@ for the full command-to-token matrix, use the [`auth-matrix`](https://kagi.micr.
 | --- | --- |
 | `kagi search` | search Kagi with `json` by default, or render as `toon`, `pretty`, `compact`, `markdown`, or `csv` |
 | `kagi agent` | print the embedded agent skill guide for choosing commands, auth paths, and output formats |
-| `kagi skills` | list and load embedded, version-matched agent skills with `skills get kagi` as the agent starting point |
+| `kagi skills` | list and load embedded, version-matched agent skills with `skills get kagi-usage` as the agent starting point |
 | `kagi batch` | run multiple searches in parallel with JSON, TOON, compact, pretty, markdown, or csv output and shared filters |
 | `kagi auth` | launch the auth wizard, or inspect, validate, and save credentials |
 | `kagi completion` | generate or install shell completions for bash, zsh, fish, or PowerShell |
diff --git a/docs/commands/skills.mdx b/docs/commands/skills.mdx
index ff6fe76..4018181 100644
--- a/docs/commands/skills.mdx
+++ b/docs/commands/skills.mdx
@@ -10,14 +10,15 @@ Load embedded, version-matched usage guidance for agents and automation.
 The top-level help points agents here first:
 
 ```bash
-kagi skills get kagi
+kagi skills get kagi-usage
 ```
 
 Skills ship with the installed CLI binary, so their command examples and
 credential guidance match that exact `kagi` version.
 
-The `kagi` skill is a router. It points to focused workflow skills so agents do
-not need to load the full command surface for every task.
+The `kagi-usage` skill covers search and page content directly, and points to
+the focused workflow skills so agents do not need to load the full command
+surface for every task.
 
 ## Synopsis
 
@@ -29,7 +30,7 @@ kagi skills path [NAME]
 kagi agent
 ```
 
-`kagi agent` is a convenience alias for `kagi skills get kagi`.
+`kagi agent` is a convenience alias for `kagi skills get kagi-usage`.
 
 ## `kagi skills`
 
@@ -46,33 +47,33 @@ kagi skills list
 Print an embedded skill guide to stdout with frontmatter stripped.
 
 ```bash
-kagi skills get kagi
+kagi skills get kagi-usage
 ```
 
 Available skills:
 
 | Skill | Use for |
 | --- | --- |
-| `kagi` | Route an unclear Kagi task |
-| `kagi-research` | Search, Quick Answer, News, source discovery, and verification |
-| `kagi-content` | Extraction, summarization, page questions, and translation |
+| `kagi-usage` | Search, page content, and routing for any other Kagi task |
+| `kagi-ai` | Paid AI reasoning, coding, research, file analysis, and model selection |
 | `kagi-assistant` | Conversations, threads, attachments, and custom assistants |
 | `kagi-monitoring` | Batch, watch, notifications, history, caching, and MCP |
 | `kagi-account-config` | Auth, profiles, lenses, bangs, redirects, and site preferences |
 
-Load the router when the workflow is unclear, or load a focused skill directly:
+Load `kagi-usage` for search and page work, or load a focused skill directly:
 
 ```bash
-kagi skills get kagi
-kagi skills get kagi-research
-kagi skills get kagi-content
+kagi skills get kagi-usage
+kagi skills get kagi-ai
+kagi skills get kagi-assistant
+kagi skills get kagi-monitoring
 ```
 
 `--full` is accepted for Harbor-style compatibility. The current embedded
 skills have no separate references, so `--full` prints the same body as `get`.
 
 ```bash
-kagi skills get kagi --full
+kagi skills get kagi-usage --full
 ```
 
 ## `kagi skills path`
@@ -81,8 +82,8 @@ Print the embedded skill locator.
 
 ```bash
 kagi skills path
-kagi skills path kagi
-kagi skills path kagi-research
+kagi skills path kagi-usage
+kagi skills path kagi-assistant
 ```
 
 Because Kagi embeds skills into the Rust binary instead of installing a
diff --git a/skills/kagi-account-config/SKILL.md b/skills/kagi-account-config/SKILL.md
index 5f078c4..1d3a476 100644
--- a/skills/kagi-account-config/SKILL.md
+++ b/skills/kagi-account-config/SKILL.md
@@ -9,7 +9,7 @@ allowed-tools: Bash(kagi:*)
 Inspect current state before changing it. Never print, echo, or paste stored
 credential values into output.
 
-## Credential Model
+## Credential model
 
 | Credential | Main capabilities |
 | --- | --- |
@@ -19,9 +19,9 @@ credential values into output.
 
 Environment variables override `.kagi.toml`. `--profile NAME` selects a named
 profile. Base search prefers session auth when both search credentials are
-available unless `preferred_auth = "api"` is configured.
+available, unless `preferred_auth = "api"` is configured.
 
-## Inspect And Verify Auth
+## Inspect and verify auth
 
 ```bash
 kagi auth status
@@ -38,7 +38,7 @@ kagi auth
 kagi auth set --session-token "https://kagi.com/search?token=..."
 ```
 
-Do not pass secrets on a shared shell command line. Prefer the interactive
+Never pass secrets on a shared shell command line. Prefer the interactive
 wizard or protected environment injection in shared environments.
 
 ## Profiles
@@ -50,9 +50,9 @@ kagi --profile work auth status
 kagi --profile personal search "query" --format toon
 ```
 
-Verify the selected profile before any setting mutation.
+Verify the selected profile before changing any setting.
 
-## Search Settings
+## Search settings
 
 Inspect before creating, updating, or deleting:
 
@@ -73,10 +73,10 @@ kagi bang custom create \
   --template "https://docs.rs/releases/search?query=%s"
 ```
 
-Use each command's `--help` for mutation-specific flags. Do not delete or
-overwrite account settings without explicit user intent.
+Use each command's `--help` for the flags a mutation needs. Never delete or
+overwrite an account setting unless the user asked for it.
 
-## Custom Assistants
+## Custom assistants
 
 ```bash
 kagi assistant custom list
@@ -86,12 +86,12 @@ kagi assistant custom get "Researcher"
 Conversation and thread use belongs in `kagi-assistant`. This skill owns the
 account-level configuration of custom assistants.
 
-## Completion Criteria
+## Completion criteria
 
 Configuration work is complete when:
 
-- the active profile and credential source are known;
+- you know the active profile and credential source;
 - the requested setting is visible through its read command;
 - no secret value appears in output or persisted scripts;
 - unrelated account settings remain unchanged; and
-- destructive changes were explicitly requested.
+- you made a destructive change only when the user asked for it.
diff --git a/skills/kagi-ai/SKILL.md b/skills/kagi-ai/SKILL.md
new file mode 100644
index 0000000..c783675
--- /dev/null
+++ b/skills/kagi-ai/SKILL.md
@@ -0,0 +1,126 @@
+---
+name: kagi-ai
+description: "Use Kagi's paid AI allowance for reasoning, coding, research, file analysis, structured generation, or model comparison. Use when a task should spend Assistant value instead of only searching the web or transforming one known page."
+allowed-tools: Bash(kagi:*) Skill(kagi:*)
+---
+
+# Kagi AI
+
+Spend the Assistant allowance on reasoning and synthesis. When the active plan
+includes unlimited Summarizer or Translate usage, use those tools for direct
+transformations and reserve the allowance for harder work.
+
+## Choose the lane
+
+| Need | Lane |
+| --- | --- |
+| Find or verify sources | Load `kagi-usage` and search |
+| Summarize one URL or text | `kagi summarize --subscriber` |
+| Translate supplied text | `kagi translate` |
+| Get a fast cited answer | `kagi quick` |
+| Reason, code, synthesize, or analyze files | `kagi assistant` |
+| Run multi-step research or tool use | An account-visible Research assistant |
+| Manage threads or custom assistants | Load `kagi-assistant` |
+
+The Assistant allowance is separate from API credit. A plan's monthly value
+sets its token allowance across Assistant models. Premium models, web access,
+large files, and long thread histories consume more of that allowance.
+
+## Start with live account state
+
+```bash
+kagi auth status
+kagi assistant models
+kagi assistant custom list
+```
+
+Use only model slugs and assistant names returned by the account. When the
+model catalog is empty, use the account default instead of guessing a slug.
+Kagi Quick and Kagi Research are usable through `--assistant` only when they
+appear in the assistant listing.
+
+## Use Assistant deliberately
+
+Use a closed model call for reasoning that needs no current information:
+
+```bash
+kagi assistant \
+  "find the weakest assumption in this design and propose a simpler one" \
+  --no-web-access \
+  --format markdown
+```
+
+Enable web access when the answer needs current sources:
+
+```bash
+kagi assistant \
+  "compare these database options for this workload and cite current limits" \
+  --web-access \
+  --format markdown
+```
+
+Choose a live model explicitly only when its cost, speed, or capability matters:
+
+```bash
+kagi assistant \
+  "review this migration plan for data-loss risks" \
+  --model MODEL_SLUG \
+  --no-web-access \
+  --format markdown
+```
+
+For a model comparison, send the same bounded prompt to each model in a fresh
+thread. Compare correctness, latency, and usefulness, not prose style alone.
+
+## Analyze files
+
+Attach only the files needed for the answer. Assistant accepts documents,
+spreadsheets, images, and audio; one upload can be at most 30 MB.
+
+```bash
+kagi assistant \
+  --attach ./report.pdf \
+  "extract the decision, evidence, risks, and unresolved questions" \
+  --no-web-access \
+  --format markdown
+```
+
+Uploaded content remains in the thread context. Start a new thread when the
+next task should not inherit it.
+
+## Require structured output
+
+Use a built-in contract for reusable decisions, plans, and checklists:
+
+```bash
+kagi assistant \
+  "decide whether this service is ready to launch" \
+  --contract decision \
+  --format json
+```
+
+Use `--contract-file PATH` when the consumer needs a different small JSON
+shape. Contract mode validates the response and makes one repair attempt.
+
+## Budget discipline
+
+1. Use subscriber Summarizer or Translate for direct transformations when the
+   active plan includes them.
+2. Use the account default or a lower-cost model for simple work.
+3. Disable web access for closed-context work.
+4. Keep prompts specific and attach only relevant files.
+5. Start a new thread for an unrelated task.
+6. Reserve Research assistants and premium models for work that needs their
+   extra tools or reasoning depth.
+
+## Completion criteria
+
+AI work is complete when:
+
+- the selected lane matches the task;
+- every explicit model or assistant came from live account discovery;
+- web access and attachments are limited to what the answer needs;
+- structured output passes its contract when a machine will consume it;
+- the response answers the task instead of spending allowance on avoidable
+  transformations; and
+- you changed or deleted account state only when the user asked.
diff --git a/skills/kagi-assistant/SKILL.md b/skills/kagi-assistant/SKILL.md
index 5bf9a05..fcd379b 100644
--- a/skills/kagi-assistant/SKILL.md
+++ b/skills/kagi-assistant/SKILL.md
@@ -6,12 +6,13 @@ allowed-tools: Bash(kagi:*)
 
 # Kagi Assistant
 
-Use Assistant for conversational synthesis and continuity. Use
-`kagi-research` for source discovery and `kagi-content` for one known page.
+Use Assistant for conversation mechanics and continuity. Use `kagi-ai` to
+choose when to spend AI allowance, and `kagi-usage` for direct search or page
+work.
 
 Assistant commands require `KAGI_SESSION_TOKEN`.
 
-## Start A Conversation
+## Start a conversation
 
 ```bash
 kagi auth status
@@ -19,11 +20,11 @@ kagi assistant "draft a migration checklist" --format markdown
 kagi assistant "compare these options" --format toon
 ```
 
-State the desired deliverable, constraints, and audience in the prompt. Do not
-hide required output structure in follow-up shell processing when the Assistant
-can produce it directly.
+State the deliverable, constraints, and audience in the prompt. Never rebuild
+the output structure in follow-up shell processing when the Assistant can
+produce it directly.
 
-## Continue A Thread
+## Continue a thread
 
 ```bash
 kagi assistant --thread-id THREAD_ID "add rollback steps"
@@ -35,13 +36,13 @@ kagi assistant thread export THREAD_ID --format markdown
 Reuse a thread only when its prior context still belongs to the task. Start a
 new thread when stale context could bias the answer.
 
-Delete a thread only when the user asked for deletion:
+Delete a thread only when the user asks for it:
 
 ```bash
 kagi assistant thread delete THREAD_ID
 ```
 
-## Attach Local Context
+## Attach local context
 
 ```bash
 kagi assistant \
@@ -50,10 +51,10 @@ kagi assistant \
   --format markdown
 ```
 
-Attach only files needed for the task. Mention the file's role in the prompt so
-the Assistant knows whether it is evidence, a template, or background.
+Attach only files the task needs. Name the file's role in the prompt so the
+Assistant knows whether it is evidence, a template, or background.
 
-## Stream Responses
+## Stream responses
 
 Use human streaming for interactive terminal work. Use structured streaming
 when another program consumes events.
@@ -66,9 +67,9 @@ kagi assistant \
   "produce a release checklist"
 ```
 
-Do not parse human-oriented streaming output as a machine contract.
+Never parse human streaming output as a machine contract.
 
-## Custom Assistants
+## Custom assistants
 
 ```bash
 kagi assistant custom list
@@ -82,12 +83,12 @@ kagi assistant custom create \
 Inspect an existing custom assistant before changing the workflow around it.
 Account configuration belongs in `kagi-account-config`.
 
-## Completion Criteria
+## Completion criteria
 
 Assistant work is complete when:
 
-- the correct thread or a clean new thread was used;
-- required files and constraints were supplied;
-- output format matches its consumer;
+- you used the correct thread or a clean new one;
+- you supplied the required files and constraints;
+- the output format matches its consumer;
 - the final response answers the current prompt, not stale thread context; and
-- destructive thread actions occurred only with explicit user intent.
+- you deleted a thread only when the user asked for it.
diff --git a/skills/kagi-content/SKILL.md b/skills/kagi-content/SKILL.md
deleted file mode 100644
index da4b3b3..0000000
--- a/skills/kagi-content/SKILL.md
+++ /dev/null
@@ -1,92 +0,0 @@
----
-name: kagi-content
-description: Read, extract, summarize, question, or translate a URL or supplied text with Kagi. Use when the source is already known and the user needs its readable content, key points, an answer about the page, or a translation.
-allowed-tools: Bash(kagi:*)
----
-
-# Kagi Content
-
-Use the narrowest content command. Do not run a broad web search when the user
-already supplied the source.
-
-## Choose The Command
-
-| Outcome | Command | Credential |
-| --- | --- | --- |
-| Full readable page markdown | `kagi extract` | `KAGI_API_KEY` |
-| Summary of a URL or text | `kagi summarize` | session or legacy API token |
-| Answer about one page | `kagi ask-page` | session token |
-| Translation | `kagi translate` | session token |
-
-Run `kagi auth status` before selecting between subscriber and API
-summarization.
-
-## Extract A Page
-
-Use `extract` when downstream work needs the article body rather than a
-summary.
-
-```bash
-kagi extract "https://example.com/article"
-```
-
-Prefer extraction over browser scraping for readable main content. Extraction
-requires the current `KAGI_API_KEY`, not the legacy `KAGI_API_TOKEN`.
-
-## Summarize
-
-Prefer subscriber mode when a session token is available.
-
-```bash
-kagi summarize --subscriber --url "https://example.com/article"
-kagi summarize --subscriber --text "long text"
-kagi summarize --subscriber \
-  --url "https://example.com/article" \
-  --summary-type keypoints
-```
-
-Use the public API mode only when the legacy API token is configured or its
-engine controls are required.
-
-```bash
-kagi summarize --url "https://example.com/article"
-kagi summarize --text "long text" --engine cecil
-```
-
-Do not summarize content that must be quoted or checked precisely. Extract it
-and inspect the relevant passage instead.
-
-## Ask One Page
-
-Use `ask-page` for a focused question whose evidence should come from one URL.
-
-```bash
-kagi ask-page \
-  "https://example.com/article" \
-  "what evidence supports the main claim?" \
-  --format markdown
-```
-
-If the question requires outside evidence or comparison, switch to
-`kagi-research`.
-
-## Translate
-
-```bash
-kagi translate \
-  --text "Bonjour tout le monde" \
-  --target-language EN
-```
-
-Preserve code, URLs, product names, and other text that should not be
-translated. State the desired target language explicitly.
-
-## Completion Criteria
-
-Content work is complete when:
-
-- the command matches the requested transformation;
-- the output stays grounded in the supplied URL or text;
-- quotes remain distinguishable from summaries;
-- the requested language and format are preserved; and
-- any source-access failure is reported rather than filled with guesses.
diff --git a/skills/kagi-monitoring/SKILL.md b/skills/kagi-monitoring/SKILL.md
index 00b2fdf..08fd1e0 100644
--- a/skills/kagi-monitoring/SKILL.md
+++ b/skills/kagi-monitoring/SKILL.md
@@ -4,12 +4,12 @@ description: Automate repeated Kagi queries with batch search, watches, notifica
 allowed-tools: Bash(kagi:*)
 ---
 
-# Kagi Monitoring And Automation
+# Kagi Monitoring and Automation
 
 Build repeatable workflows from structured Kagi output. Keep stdout
 machine-readable and send progress or diagnostics to stderr.
 
-## Choose The Workflow
+## Choose the workflow
 
 | Need | Command |
 | --- | --- |
@@ -21,7 +21,7 @@ machine-readable and send progress or diagnostics to stderr.
 
 Run `kagi auth status` before automating authenticated commands.
 
-## Batch Search
+## Batch search
 
 ```bash
 kagi batch "rust" "zig" "go" --format toon --limit 3
@@ -29,28 +29,28 @@ printf 'rust\nzig\ngo\n' | kagi batch --format compact
 ```
 
 Use argument queries for short fixed sets and stdin for generated lists. Keep
-the query list as the recoverable input so failed runs can be repeated.
+the query list as the recoverable input so a failed run can be repeated.
 
-## Watch Search Results
+## Watch search results
 
 ```bash
 kagi watch "site:example.com release notes" --interval 300
 ```
 
 Define what counts as a meaningful change before starting a long-running
-watch. Use a descriptive query and an interval that matches the source's likely
-update rate.
+watch. Use a descriptive query and an interval that matches how often the
+source updates.
 
 ## Notifications
 
 Use `kagi notify --help` to select the configured delivery target, then connect
-it to a watch or batch result. Do not put secrets in notification text or
+it to a watch or batch result. Never put secrets in notification text or
 command history.
 
-## History And Caching
+## History and caching
 
 Use `kagi history` to inspect local command records before duplicating work.
-Use `--local-cache` only for calls where stale data is acceptable. Do not cache
+Use `--local-cache` only for calls where stale data is acceptable. Never cache
 research that must reflect a current price, release, outage, or policy.
 
 ## MCP
@@ -61,24 +61,24 @@ Use the stdio server when another agent host needs Kagi tools:
 kagi mcp
 ```
 
-Keep MCP stdout reserved for protocol messages. Treat logs written to stdout as
-a protocol-breaking bug.
+Keep MCP stdout reserved for protocol messages. Treat a log written to stdout
+as a protocol-breaking bug.
 
-## Automation Rules
+## Automation rules
 
 - Prefer `json`, `compact`, or `toon` over `pretty`.
-- Preserve nonzero exit codes instead of converting failure into empty output.
-- Keep credentials in Kagi auth storage or environment variables, not scripts.
+- Preserve nonzero exit codes instead of turning failure into empty output.
+- Keep credentials in Kagi auth storage or environment variables, never scripts.
 - Limit concurrency to what the endpoint and account can sustain.
 - Record enough input to reproduce a failed item.
 - Use a process supervisor or scheduler for long-running watches.
 
-## Completion Criteria
+## Completion criteria
 
 Automation is complete when:
 
 - inputs and outputs have stable machine-readable shapes;
 - failures remain observable and retryable;
-- credential values do not appear in scripts or logs;
+- credential values never appear in scripts or logs;
 - cache and polling choices match the freshness requirement; and
 - long-running processes have a clear owner and stop condition.
diff --git a/skills/kagi-research/SKILL.md b/skills/kagi-research/SKILL.md
deleted file mode 100644
index acfeac6..0000000
--- a/skills/kagi-research/SKILL.md
+++ /dev/null
@@ -1,94 +0,0 @@
----
-name: kagi-research
-description: Research, compare, verify, or discover information with Kagi Search, Quick Answer, News, and Small Web. Use when the user needs source discovery, current facts, multiple perspectives, or a cited answer rather than one known page.
-allowed-tools: Bash(kagi:*)
----
-
-# Kagi Research
-
-Use Kagi to discover sources, narrow the question, and return evidence that
-matches the user's requested depth.
-
-## Choose The Research Lane
-
-| Need | Command |
-| --- | --- |
-| Direct answer with references | `kagi quick` |
-| Ranked web results | `kagi search` |
-| Search and summarize top pages | `kagi search --follow N` |
-| Current news stories | `kagi news` or `kagi search --news` |
-| Independent personal sites | `kagi smallweb` |
-| Several independent queries | `kagi batch` |
-
-`quick`, search filters, and the subscriber search path require session auth.
-The current Search API uses `KAGI_API_KEY`. Public `news` and `smallweb` need no
-credentials.
-
-## Workflow
-
-1. Define the claim or question that needs evidence.
-2. Run `kagi auth status`.
-3. Start with one broad search or `quick` for a bounded factual question.
-4. Split the topic into distinct queries when one query cannot cover it.
-5. Prefer primary sources and direct evidence.
-6. Use `--follow` only when summaries of the top result pages are useful.
-7. Report source disagreement instead of flattening it.
-
-```bash
-kagi auth status
-kagi search "rust async cancellation" --format toon --limit 5
-kagi quick "what is the current Kagi Search API endpoint?" --format markdown
-```
-
-## Search Controls
-
-```bash
-kagi search "query" --region us --format toon --limit 10
-kagi search "query" --time month --order recency --format json
-kagi search "query" --snap reddit --format toon
-kagi search "query" --lens 2 --format toon
-kagi search "query" --follow 3 --format markdown
-```
-
-Session-only controls include `--lens`, `--time`, `--order`, `--verbatim`,
-personalization flags, and News vertical search. Use `--local-cache` only when
-stale results are acceptable.
-
-## Query Design
-
-- Quote exact error messages or disputed phrases.
-- Add official domains with `site:` when a primary source is known.
-- Use native-language query terms for regional or multilingual research.
-- Separate discovery, counterevidence, and recency into different queries.
-- Use `--region` as a result bias, not as a language guarantee.
-
-```bash
-kagi batch \
-  "topic official documentation" \
-  "topic limitations" \
-  "topic independent review" \
-  --format toon --limit 5
-```
-
-## News And Discovery
-
-```bash
-kagi news --category tech --limit 10
-kagi news --list-categories
-kagi news --chaos
-kagi search "open source ai" --news --format toon
-kagi smallweb
-```
-
-Use `news` for the public Kagi News feed. Use `search --news` when the query must
-filter the News vertical.
-
-## Completion Criteria
-
-Research is complete when:
-
-- the answer addresses the exact question;
-- important claims trace to source URLs;
-- source quality and recency match the claim;
-- disagreements or uncertainty are explicit; and
-- the output uses the requested format and depth.
diff --git a/skills/kagi-usage/SKILL.md b/skills/kagi-usage/SKILL.md
new file mode 100644
index 0000000..3fdbd8a
--- /dev/null
+++ b/skills/kagi-usage/SKILL.md
@@ -0,0 +1,205 @@
+---
+name: kagi-usage
+description: "Search the web, extract or summarize a page, translate it, or check the news with the kagi CLI instead of WebSearch or WebFetch. Load this first for any Kagi task: it covers search and page content, and routes Assistant, automation, and account work to the right skill."
+allowed-tools: Bash(kagi:*) Skill(kagi:*)
+---
+
+# Kagi CLI
+
+Search the web and read pages with `kagi`. Route every other Kagi task to the
+skill named below.
+
+## Route the task
+
+| User intent | Where |
+| --- | --- |
+| Find, compare, or verify information from the web | Search the web (below) |
+| Read, summarize, question, or translate a known page | Work with one page (below) |
+| Use paid AI for reasoning, coding, research, or files | `kagi-ai` |
+| Continue a conversation or manage Assistant state | `kagi-assistant` |
+| Run searches in bulk or monitor changes over time | `kagi-monitoring` |
+| Configure credentials or account search behavior | `kagi-account-config` |
+
+Load the skill when it appears in the available skills listing. Otherwise run
+`kagi skills get <name>` for the same guidance.
+
+```bash
+kagi skills list
+kagi skills get kagi-assistant
+```
+
+## Search the web
+
+Discover sources, narrow the question, and return evidence that matches the
+depth the user asked for.
+
+| Need | Command |
+| --- | --- |
+| Direct answer with references | `kagi quick` |
+| Ranked web results | `kagi search` |
+| Search and summarize top pages | `kagi search --follow N` |
+| Current news stories | `kagi news` or `kagi search --news` |
+| Independent personal sites | `kagi smallweb` |
+| Several independent queries | `kagi batch` |
+
+`quick`, session-only search controls, and the subscriber search path require
+session auth.
+The current Search API uses `KAGI_API_KEY`. Public `news` and `smallweb` need no
+credentials.
+
+1. Define the claim or question that needs evidence.
+2. Start with one broad search, or `quick` for a bounded factual question.
+3. Split the topic into distinct queries when one query cannot cover it.
+4. Prefer primary sources and direct evidence.
+5. Use `--follow` only when the user needs the top result pages summarized.
+6. Report source disagreement instead of flattening it.
+
+```bash
+kagi search "rust async cancellation" --format toon --limit 5
+kagi quick "what is the current Kagi Search API endpoint?" --format markdown
+```
+
+### Search controls
+
+```bash
+kagi search "query" --region us --format toon --limit 10
+kagi search "query" --time month --order recency --format json
+kagi search "query" --snap reddit --format toon
+kagi search "query" --lens 2 --format toon
+kagi search "query" --follow 3 --format markdown
+```
+
+Session-only controls include `--lens`, `--time`, `--order`, `--verbatim`,
+personalization flags, and News vertical search. Use `--local-cache` only when
+stale results are acceptable.
+
+### Query design
+
+- Quote exact error messages or disputed phrases.
+- Add official domains with `site:` when a primary source is known.
+- Use native-language query terms for regional or multilingual research.
+- Separate discovery, counterevidence, and recency into different queries.
+- Use `--region` as a result bias, never as a language guarantee.
+
+```bash
+kagi batch \
+  "topic official documentation" \
+  "topic limitations" \
+  "topic independent review" \
+  --format toon --limit 5
+```
+
+### News and discovery
+
+```bash
+kagi news --category tech --limit 10
+kagi news --list-categories
+kagi news --chaos
+kagi search "open source ai" --news --format toon
+kagi smallweb
+```
+
+Use `news` for the public Kagi News feed. Use `search --news` when the query must
+filter the News vertical.
+
+## Work with one page
+
+Work from the source the user supplied. Search the web only when the task needs
+evidence that page cannot give.
+
+| Outcome | Command | Credential |
+| --- | --- | --- |
+| Full readable page markdown | `kagi extract` | `KAGI_API_KEY` |
+| Summary of a URL or text | `kagi summarize` | session or legacy API token |
+| Answer about one page | `kagi ask-page` | session token |
+| Translation | `kagi translate` | session token |
+
+### Extract a page
+
+Use `extract` when downstream work needs the article body rather than a
+summary. Prefer `extract` over browser scraping for readable main content. It
+requires the current `KAGI_API_KEY`, never the legacy `KAGI_API_TOKEN`.
+
+```bash
+kagi extract "https://example.com/article"
+```
+
+### Summarize
+
+Prefer subscriber mode when a session token is available.
+
+```bash
+kagi summarize --subscriber --url "https://example.com/article"
+kagi summarize --subscriber --text "long text"
+kagi summarize --subscriber \
+  --url "https://example.com/article" \
+  --summary-type keypoints
+```
+
+Use the public API mode only when the legacy API token is configured or the
+task needs its engine controls.
+
+```bash
+kagi summarize --url "https://example.com/article"
+kagi summarize --text "long text" --engine cecil
+```
+
+Never summarize content that must be quoted or checked precisely. Extract it
+and inspect the relevant passage instead.
+
+### Ask one page
+
+Use `ask-page` for a focused question whose evidence should come from one URL.
+
+```bash
+kagi ask-page \
+  "https://example.com/article" \
+  "what evidence supports the main claim?" \
+  --format markdown
+```
+
+When the question needs outside evidence or comparison, search the web instead.
+
+### Translate
+
+```bash
+kagi translate \
+  --text "Bonjour tout le monde" \
+  --target-language EN
+```
+
+Preserve code, URLs, product names, and other text that must stay untranslated.
+State the target language explicitly.
+
+## Shared rules
+
+1. Run `kagi auth status` before treating a command as unavailable.
+2. Choose the narrowest command that produces the requested outcome.
+3. Use `--format json` for programmatic parsing.
+4. Use `--format toon` for compact LLM context.
+5. Use `--format markdown` for prose that a person will read.
+6. Run `kagi auth check` when credentials exist but a request fails.
+7. Never print credential values.
+
+```bash
+kagi auth status
+kagi auth check
+```
+
+## Completion criteria
+
+Web research is complete when:
+
+- the answer addresses the exact question;
+- important claims trace to source URLs;
+- source quality and recency match the claim;
+- disagreement and uncertainty appear in the output; and
+- the output uses the requested format and depth.
+
+Page work is complete when:
+
+- the command matches the requested transformation;
+- the output stays grounded in the supplied URL or text;
+- quotes remain distinguishable from summaries;
+- the output keeps the requested language and format; and
+- you report a source-access failure instead of guessing past it.
diff --git a/skills/kagi/SKILL.md b/skills/kagi/SKILL.md
deleted file mode 100644
index 9fc78fe..0000000
--- a/skills/kagi/SKILL.md
+++ /dev/null
@@ -1,50 +0,0 @@
----
-name: kagi
-description: Route Kagi CLI work to version-matched skills for research, page content, Assistant, monitoring, or account configuration. Use when a task involves Kagi but the correct command or specialized Kagi skill is not yet clear.
-allowed-tools: Bash(kagi:*)
----
-
-# Kagi CLI
-
-Use this skill as the entry point for `kagi`. Route the task to the narrowest
-workflow skill before running substantive commands.
-
-## Route The Task
-
-| User intent | Load |
-| --- | --- |
-| Find, compare, or verify information from the web | `kagi skills get kagi-research` |
-| Read, summarize, question, or translate a page | `kagi skills get kagi-content` |
-| Start or continue an Assistant conversation | `kagi skills get kagi-assistant` |
-| Run searches in bulk or monitor changes over time | `kagi skills get kagi-monitoring` |
-| Configure credentials or account search behavior | `kagi skills get kagi-account-config` |
-
-```bash
-kagi skills list
-kagi skills get kagi-research
-```
-
-If one task crosses boundaries, load only the skills needed for the next action.
-For example, research first with `kagi-research`, then load `kagi-content` only
-when a result page needs closer reading.
-
-## Shared Rules
-
-1. Run `kagi auth status` before deciding that a command is available.
-2. Choose the narrowest command for the requested outcome.
-3. Use `--format json` for programmatic parsing.
-4. Use `--format toon` for compact LLM context.
-5. Use `--format markdown` for prose that a person will read.
-6. Run `kagi auth check` when credentials exist but a request fails.
-7. Never print credential values.
-
-```bash
-kagi auth status
-kagi auth check
-```
-
-## Completion Criteria
-
-Routing is complete when the narrowest workflow skill has been loaded and its
-completion criteria have been satisfied. Do not keep this router in context
-when a specialized skill contains all remaining guidance.
diff --git a/src/agent.rs b/src/agent.rs
index d682f4f..e281e13 100644
--- a/src/agent.rs
+++ b/src/agent.rs
@@ -8,11 +8,10 @@ pub struct EmbeddedSkill {
 }
 
 /// Core skill name used by `kagi skills`.
-pub const KAGI_SKILL: &str = "kagi";
+pub const KAGI_SKILL: &str = "kagi-usage";
 
-const KAGI_SKILL_SOURCE: &str = include_str!("../skills/kagi/SKILL.md");
-const KAGI_RESEARCH_SKILL_SOURCE: &str = include_str!("../skills/kagi-research/SKILL.md");
-const KAGI_CONTENT_SKILL_SOURCE: &str = include_str!("../skills/kagi-content/SKILL.md");
+const KAGI_SKILL_SOURCE: &str = include_str!("../skills/kagi-usage/SKILL.md");
+const KAGI_AI_SKILL_SOURCE: &str = include_str!("../skills/kagi-ai/SKILL.md");
 const KAGI_ASSISTANT_SKILL_SOURCE: &str = include_str!("../skills/kagi-assistant/SKILL.md");
 const KAGI_MONITORING_SKILL_SOURCE: &str = include_str!("../skills/kagi-monitoring/SKILL.md");
 const KAGI_ACCOUNT_CONFIG_SKILL_SOURCE: &str =
@@ -21,18 +20,13 @@ const KAGI_ACCOUNT_CONFIG_SKILL_SOURCE: &str =
 const SKILLS: &[EmbeddedSkill] = &[
     EmbeddedSkill {
         name: KAGI_SKILL,
-        description: "Route Kagi CLI tasks to the right embedded workflow skill",
+        description: "Search the web, read pages, and route other Kagi tasks to their skill",
         source: KAGI_SKILL_SOURCE,
     },
     EmbeddedSkill {
-        name: "kagi-research",
-        description: "Research a topic with Kagi Search, Quick Answer, News, and source follow-up",
-        source: KAGI_RESEARCH_SKILL_SOURCE,
-    },
-    EmbeddedSkill {
-        name: "kagi-content",
-        description: "Extract, summarize, question, or translate web page content",
-        source: KAGI_CONTENT_SKILL_SOURCE,
+        name: "kagi-ai",
+        description: "Spend Kagi AI allowance on reasoning, research, coding, and file analysis",
+        source: KAGI_AI_SKILL_SOURCE,
     },
     EmbeddedSkill {
         name: "kagi-assistant",
@@ -144,9 +138,8 @@ mod tests {
         assert_eq!(
             names,
             [
-                "kagi",
-                "kagi-research",
-                "kagi-content",
+                "kagi-usage",
+                "kagi-ai",
                 "kagi-assistant",
                 "kagi-monitoring",
                 "kagi-account-config",
@@ -163,5 +156,9 @@ mod tests {
                 "{name} should expose a frontmatter-free skill body"
             );
         }
+
+        let ai = skill_content("kagi-ai").expect("kagi-ai should be embedded");
+        assert!(ai.contains("Assistant allowance"));
+        assert!(ai.contains("kagi assistant"));
     }
 }
diff --git a/src/cli.rs b/src/cli.rs
index a89c355..523de43 100644
--- a/src/cli.rs
+++ b/src/cli.rs
@@ -258,14 +258,14 @@ Features:
 - Full Kagi API coverage with session token support
 
 Agent usage:
-    kagi skills get kagi
+    kagi skills get kagi-usage
 
     Skills ship with the CLI and are always version-matched. They include
     workflow patterns, credential guidance, and copy-paste examples. Prefer
     this over guessing commands from flag docs alone.
 
     skills [list]            List available skills
-    skills get kagi          Core CLI usage guide
+    skills get kagi-usage    Core CLI usage guide
     skills get <name>        Load a specialized skill
     skills path [name]       Print the embedded skill locator",
     propagate_version = true
diff --git a/tests/integration-cli.rs b/tests/integration-cli.rs
index 63cf1f1..57916f4 100644
--- a/tests/integration-cli.rs
+++ b/tests/integration-cli.rs
@@ -128,8 +128,8 @@ fn help_points_agents_to_agent_guide() {
         "expected agent help section, got:\n{stdout}"
     );
     assert!(
-        stdout.contains("kagi skills get kagi"),
-        "expected help to mention kagi skills get kagi, got:\n{stdout}"
+        stdout.contains("kagi skills get kagi-usage"),
+        "expected help to mention kagi skills get kagi-usage, got:\n{stdout}"
     );
     assert!(
         stdout.contains("skills [list]"),
@@ -241,7 +241,7 @@ fn agent_command_prints_embedded_skill_guide_without_auth() {
 #[test]
 fn skills_get_prints_core_guide_without_auth() {
     let tempdir = TempDir::new().expect("tempdir");
-    let output = run_kagi(&["skills", "get", "kagi"], &[], tempdir.path());
+    let output = run_kagi(&["skills", "get", "kagi-usage"], &[], tempdir.path());
 
     assert_success(&output);
     let stdout = String::from_utf8_lossy(&output.stdout);
@@ -250,7 +250,7 @@ fn skills_get_prints_core_guide_without_auth() {
         "expected markdown skill guide, got:\n{stdout}"
     );
     assert!(
-        stdout.contains("kagi skills get kagi"),
+        stdout.contains("kagi skills get <name>"),
         "expected skills command guidance, got:\n{stdout}"
     );
 }
@@ -263,9 +263,8 @@ fn skills_list_and_path_are_auth_free() {
     assert_success(&list);
     let stdout = String::from_utf8_lossy(&list.stdout);
     for skill in [
-        "kagi",
-        "kagi-research",
-        "kagi-content",
+        "kagi-usage",
+        "kagi-ai",
         "kagi-assistant",
         "kagi-monitoring",
         "kagi-account-config",
@@ -283,18 +282,22 @@ fn skills_list_and_path_are_auth_free() {
         "embedded://skills"
     );
 
-    let skill_path = run_kagi(&["skills", "path", "kagi"], &[], tempdir.path());
+    let skill_path = run_kagi(&["skills", "path", "kagi-usage"], &[], tempdir.path());
     assert_success(&skill_path);
     assert_eq!(
         String::from_utf8_lossy(&skill_path.stdout).trim(),
-        "embedded://skills/kagi"
+        "embedded://skills/kagi-usage"
     );
 }
 
 #[test]
 fn skills_get_full_prints_body_without_frontmatter() {
     let tempdir = TempDir::new().expect("tempdir");
-    let output = run_kagi(&["skills", "get", "kagi", "--full"], &[], tempdir.path());
+    let output = run_kagi(
+        &["skills", "get", "kagi-usage", "--full"],
+        &[],
+        tempdir.path(),
+    );
 
     assert_success(&output);
     let stdout = String::from_utf8_lossy(&output.stdout);