feat: add wt step eval for template expression evaluation

[max-sixty]

Summary

• Add experimental wt step eval subcommand that evaluates a minijinja template in the current worktree context and prints the result to stdout
• Reuses existing hook template infrastructure — all variables (branch, repo, worktree_path, etc.) and filters (hash_port, sanitize, sanitize_db) are available
• Output goes to stdout with no decoration or shell escaping, designed for scripting and shell substitution

$ wt step eval '{{ branch | hash_port }}'
16066

$ curl http://localhost:$(wt step eval '{{ branch | hash_port }}')/health

This is the first step from the IDEAS.md design doc, directly addressing the user request in #947 — "found no way to run it on a single branch."

Test plan

• 6 integration tests: basic branch, hash_port, multiple values, sanitize_db, conditionals, template error
• Help text and completion snapshots updated
• stdout guard allowlisted
• Doc pages auto-synced
• All lints pass
• 2360 tests pass (5 pre-existing shell_wrapper failures excluded)

Ref #947

This was written by Claude Code on behalf of @max-sixty

[max-sixty]

Follow-up PR for per-branch kv state: #1006

This was written by Claude Code on behalf of @max-sixty

[worktrunk-bot]

CI failed on all three platforms with two test failures caused by this PR:

1. test_builtin_step_commands_matches_clap — "eval" needs to be added to BUILTIN_STEP_COMMANDS in src/commands/alias.rs (the list that prevents aliases from silently conflicting with built-in step commands).

2. test_eval_template_error — snapshot mismatch; the committed snapshot has env var names that differ from what CI produces. Needs regeneration via cargo insta test --accept.

I can push a fix commit for both if you'd like.

CI failed: eval missing from BUILTIN_STEP_COMMANDS and snapshot mismatch in test_eval_template_error

[max-sixty]

@worktrunk-bot resolve conflicts. consider adding a --dry-run flag to show the reified template

[worktrunk-bot]

Resolved conflicts and added --dry-run flag to wt step eval.

Conflicts resolved: Three files had conflicts from main's doc format changes (badge-experimental spans, anchor links in operations list, subdoc restructuring). Took main's format and added the eval entries.

--dry-run flag: Shows all template variables and their current values, plus the expanded result — useful for debugging templates:

$ wt step eval --dry-run '{{ branch | hash_port }}'
branch=feature/auth-oauth2
worktree_path=/home/user/projects/myapp-feature-auth-oauth2
...
---
Result: 16066

Variables go to stderr, so --dry-run output doesn't interfere with piping. All 1187 integration tests + 463 unit tests pass locally. Waiting on CI.

codecov/patch failed (74.07%) — mostly from merging main, but the dry_run path in eval.rs is also untested. See follow-up comment.

[worktrunk-bot]

codecov/patch reports 74.07% — the bulk of the gap is from merging main into this branch (codecov measures all changed lines in the diff, including the merged-in code).

For the eval-specific code: the --dry-run path in src/commands/eval.rs (lines 40–47) has no test coverage. Adding a test_eval_dry_run integration test would close this gap. I can push a fix commit if you'd like.

[max-sixty]

@worktrunk-bot fix codecov

[worktrunk-bot]

Added test_eval_dry_run integration test covering the --dry-run code path in src/commands/eval.rs (lines 40–47). All CI checks pass — codecov/patch hasn't reported a new status for this commit (it may not re-report when the gap is resolved).

[worktrunk-bot]

Duplicated line.

Suggested change

	/// With `dry_run`, prints the template variables and the expanded result
	/// With `dry_run`, prints the template variables and the expanded result
	/// With `dry_run`, prints the template variables and the expanded result