Add gitcron integration: native tool + skills

[patel-lyzr]

Summary

• Add native gitcron built-in tool (src/tools/gitcron.ts) that wraps the gitcron CLI for scheduling, task management, and reminders
• Register gitcron tool in src/tools/index.ts
• Add three skills: scheduling, task-management, reminders with SKILL.md definitions
• Update agent.yaml to declare the gitcron tool and skills

Test plan

• Verify npm run build compiles without errors
• Verify gitcron tool is registered and available in agent runtime
• Verify skills are loaded from skills/ directory

Author: Khush Patel (khush@lyzr.ai) @patel-lyzr

[shreyas-lyzr]

The gitcron integration adds scheduling skills and a native tool. However:

1. This modifies `agent.yaml` at the repo root — that file is a template/scaffold for new agents. Adding `gitcron` as a required tool there means every new agent gets it by default, which may not be desired.
2. The skills are well-structured but this should be opt-in via the plugin system or skills marketplace, not baked into the core scaffold.

Suggest making this a plugin or installable skill pack instead of modifying the core agent template.

[shreyas-lyzr]

Apologies for the very late review @patel-lyzr.

State of play: gitcron is not in main today. The PR is technically still relevant, but a few questions before I'd merge:

1. Is gitcron the canonical scheduling tool we want bundled as a built-in? Today the codebase has src/schedules.ts and src/schedule-runner.ts doing scheduling internally. If those already cover the use cases this PR addresses, gitcron is redundant. If they cover different concerns (e.g., gitcron is for cross-agent scheduling, internal scheduling is for per-agent), then both can coexist — but the README / docs need to explain the split.
2. Built-in vs optional? Even if gitcron is the right tool, I'd lean toward shipping it as an installable skill/plugin rather than a built-in src/tools/gitcron.ts. The skill ecosystem (skills/scheduling, skills/task-management, skills/reminders) is exactly the right place for it — and you already created those skills in this PR. Worth considering whether the built-in tool wrapper adds enough on top.
3. Rebase needed regardless — two months of drift, 446 lines diff.

If gitcron is still the planned direction, please confirm and rebase; I'll merge same-day. If it makes more sense as a separate gitclaw-gitcron-skill plugin package, I'd close this PR and you could ship that independently.

Real apologies for the wait. Either direction is fine — just need a call on which.