mirror of https://github.com/outbackdingo/optimclaw.git synced 2026-08-26 23:50:17 +00:00

Files

T

[email protected]andClaude Opus 4.6 8e2349d12e feat(skills): extract ironclaw_skills crate and integrate with v2 engine

Extract the skills system into a standalone `ironclaw_skills` crate
(following the ironclaw_safety pattern) and wire it into the v2 engine
for deterministic skill selection, CodeAct code injection, and
confidence tracking.

**ironclaw_skills crate** (94 tests):
- Core types: SkillManifest, ActivationCriteria, LoadedSkill, SkillTrust
- V2 types: V2SkillMetadata, CodeSnippet, SkillMetrics, V2SkillSource
- Deterministic 4-phase selector (gating→scoring→budget→attenuation)
- apply_confidence_factor() for extracted skill scoring
- SKILL.md parser, validation/escaping, gating, registry, catalog
- Feature-gated: catalog (reqwest), registry (filesystem)

**Engine integration** (14 new tests):
- DocType::Skill with retrieval weight 0.45
- SkillSelector bridges MemoryDoc→LoadedSkill for shared scoring
- SkillTracker for usage/version/rollback confidence tracking
- System prompt injection via <skill> XML blocks
- CodeAct snippet injection via Monty NameLookup
- Skill extraction mission replaces playbook extraction
- ThreadManager.set_skill_selector() for runtime wiring

**Bridge + migration**:
- skill_migration.rs: v1 SKILL.md → v2 MemoryDoc (idempotent)
- init_engine() migrates v1 skills, builds SkillSelector
- src/skills/mod.rs → re-export shim

**E2E test** (tests/engine_v2_skill_codeact.rs):
- Full CodeAct loop: skill selected → LLM returns Python code →
  Monty executes http() → mock returns canned GitHub JSON →
  FINAL() terminates → thread completes with canned data
- GitHub SKILL.md in skills/github/ as reference implementation

Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]>

2026-03-27 16:24:07 -07:00

3.1 KiB

Raw Blame History

name, version, description, activation

name

version

description

activation

github

1.0.0

GitHub API integration via HTTP tool with automatic credential injection

keywords

exclude_keywords

patterns

GitHub API Skill

You have access to the GitHub REST API via the http tool. Credentials are automatically injected — never construct Authorization headers manually. When the URL host is api.github.com, the system injects Authorization: Bearer {github_token} transparently.

API Patterns

All endpoints use https://api.github.com as the base URL. Common headers are injected automatically.

Issues

List issues:

http(method="GET", url="https://api.github.com/repos/{owner}/{repo}/issues?state=open&sort=created&direction=desc&per_page=30")

Get single issue:

http(method="GET", url="https://api.github.com/repos/{owner}/{repo}/issues/{number}")

Create issue:

http(method="POST", url="https://api.github.com/repos/{owner}/{repo}/issues", body={"title": "...", "body": "...", "labels": ["bug"]})

Add comment:

http(method="POST", url="https://api.github.com/repos/{owner}/{repo}/issues/{number}/comments", body={"body": "..."})

Pull Requests

List PRs:

http(method="GET", url="https://api.github.com/repos/{owner}/{repo}/pulls?state=open&sort=created&direction=desc&per_page=30")

Create PR:

http(method="POST", url="https://api.github.com/repos/{owner}/{repo}/pulls", body={"title": "...", "body": "...", "head": "feature-branch", "base": "main", "draft": true})

Get PR diff:

http(method="GET", url="https://api.github.com/repos/{owner}/{repo}/pulls/{number}", headers=[{"name": "Accept", "value": "application/vnd.github.v3.diff"}])

Repository

Get repo info:

http(method="GET", url="https://api.github.com/repos/{owner}/{repo}")

List branches:

http(method="GET", url="https://api.github.com/repos/{owner}/{repo}/branches")

List recent commits:

http(method="GET", url="https://api.github.com/repos/{owner}/{repo}/commits?per_page=10")

Response Handling

GitHub returns JSON. Parse the response to extract relevant fields.
For list endpoints, check the Link header for pagination.
Rate limit: 5000 req/hour authenticated. Check X-RateLimit-Remaining header if doing bulk operations.
Errors return {"message": "..."} — always check for error responses.

Common Mistakes

Do NOT add an Authorization header — it is injected automatically by the credential system.
Always use HTTPS URLs (HTTP is blocked by the security layer).
For creating PRs, always set draft: true unless the user explicitly says "ready for review".
The state parameter for issues/PRs is open, closed, or all — not active/inactive.
Use per_page to control result count (max 100). Default is 30.

3.1 KiB Raw Blame History