Add self-improving agent pattern (.learnings/ feedback loop)

[Copilot]

Implements the self-improving-agent pattern so AI agents working in this repo capture and accumulate knowledge over time rather than losing context between sessions.

What's added

• .learnings/ — structured log directory with three template files:
  • ERRORS.md — command/tool failures with reproducibility and fix metadata
  • LEARNINGS.md — corrections, knowledge gaps, best practices (categorized)
  • FEATURE_REQUESTS.md — missing capabilities with complexity estimates
• CLAUDE.md — prepended Self-Improvement Workflow section: when to log, ID format (LRN-/ERR-/FEAT-YYYYMMDD-XXX), and promotion rules
• .github/copilot-instructions.md — new file; project conventions + self-improvement workflow for GitHub Copilot

Feedback loop

agent interaction
      │
      ▼
.learnings/{ERRORS,LEARNINGS,FEATURE_REQUESTS}.md   ← append during session
      │
      ▼ (when broadly applicable)
CLAUDE.md / copilot-instructions.md                 ← promote to permanent memory

Agents are instructed to review .learnings/ before major tasks, ensuring accumulated knowledge feeds forward into subsequent sessions.

Warning

Firewall rules blocked me from connecting to one or more addresses (expand for details)

I tried to connect to the following addresses, but was blocked by firewall rules:

• clawhub.ai
  • Triggering command: /home/REDACTED/work/_temp/ghcca-node/node/bin/node /home/REDACTED/work/_temp/ghcca-node/node/bin/node --enable-source-maps /home/REDACTED/work/_temp/copilot-developer-action-main/dist/index.js (dns block)

If you need me to access, download, or install something from one of these locations, you can either:

• Configure Actions setup steps to set up my environment, which run before the firewall is enabled
• Add the appropriate URLs or hosts to the custom allowlist in this repository's Copilot coding agent settings (admins only)

---

🔒 GitHub Advanced Security automatically protects Copilot coding agent pull requests. You can protect all pull requests by enabling Advanced Security for your repositories. Learn more about Advanced Security.