Self-Improvement Learning Log
Paste the following prompt into your AI chat to install this skill:
Please follow https://skillhub.cn/install/skillhub.md to install @user_9ac1b499/test-skill-v2.
About this skill
Problem Being Solved
AI agents often hit command failures, stale knowledge, user corrections, and inconsistent API behavior, but that context usually disappears after one session. The next task may repeat the same mistake, and developers keep explaining project conventions. This skill turns one-off fixes into searchable, trackable, reusable project knowledge.
How It Works and Boundaries
It maintains .learnings/ at the project or workspace root, using ERRORS.md, LEARNINGS.md, and FEATURE_REQUESTS.md for different event types. Entries get IDs such as ERR-YYYYMMDD-XXX. The workflow covers logging concise error summaries, tagging user corrections, linking recurring patterns with See Also, and promoting durable rules to CLAUDE.md, AGENTS.md, .github/copilot-instructions.md, or OpenClaw’s SOUL.md / TOOLS.md. It favors short redacted summaries and reproduction steps rather than full transcripts, secrets, or environment variables. It fits long-lived agent workflows, but is not intended for untrusted cross-session sharing or automatic capture of sensitive operations.
Use Cases
- After a Claude Code command exits non-zero, write a concise error summary to `.learnings/ERRORS.md` for later debugging.
- When a user says an API pattern is outdated, log a `knowledge_gap` entry in `LEARNINGS.md` and link prior related items.
- If pnpm workspace install failures keep occurring, promote the rule to `CLAUDE.md` so future agents use `pnpm install`.
- Dedupe `simplify-and-harden` candidates by `Pattern-Key`, then promote to `AGENTS.md` after three recurrences.
Best For
- Engineers maintaining Claude Code or Codex workflows who want repeated failures distilled into project rules.
- Full-stack developers writing GitHub Copilot context who need to encode pnpm and build conventions.
- OpenClaw agent operators who want shared behavioral guidelines and tool gotchas across sessions.
- Tech leads running long-lived AI-assisted projects who need reviewable learning logs and promotion workflows.
Related Skills
Turn vague requests into structured prompts with role, task, context, and constraints, outputting only usable prompts without executing the task.
A skill for game NPC and enemy AI that separates decision, steering, and pathfinding, covering FSMs, behavior trees, A*, path caching, and pitfalls.
Provides multi-turn emotional companion chat through a third-party Agent API and supports continuing sessions via session_id.
An evaluation-driven workflow for diagnosing, optimizing, validating, and logging Agent Skills based on design patterns and anti-patterns.