feat: agentic harness — CLAUDE.md, skills, permission settings - #91
Merged
Conversation
…), settings - CLAUDE.md: environment bootstrap, verified command table, architecture map, week-output conventions, skills index - .claude/skills/feature: full explore→plan→implement→test→PR pipeline with 'tests within reason' policy - .claude/skills/generate-week: offline Claude-authored themed weeks rendered via worksheet_html_renderer; teacher guide emitted by the same script - .claude/skills/test: pytest + Playwright invocation, triage, troubleshooting - .claude/settings.json: conservative permission allowlist - .gitignore: track .claude/ (except settings.local.json); ignore scripts/reference/ - scripts/generate_weather_week_series.py: canonical example produced by a cold dry-run of the generate-week skill (11 pages + embedded teacher guide) - frontend/.prettierignore: exclude Playwright artifacts so format:check passes after local e2e runs Both skills validated by cold subagent dry-runs: weather week generated clean (11 pages, all spot-checks pass); /health feature PR #90 shipped with 137 pytest + 41 e2e green. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
Instruments the repo for Claude Code sessions so recurring workflows are one command instead of re-derived context:
/featureskill — full pipeline: explore → plan → implement on a branch → tests → PR, with the tests-within-reason policy (happy-path Playwright coverage per use case, one pytest test per backend behavior, no boundary matrices)./generate-weekskill — Claude-authored themed 5-day printable weeks, fully offline (no OpenAI key), rendered viaworksheet_html_renderer; the teacher guide is emitted by the same script so gitignored output dirs stay fully reproducible./testskill — pytest + Playwright invocation patterns, failure triage, port/venv troubleshooting..claude/is now tracked (onlysettings.local.jsonignored);scripts/reference/added for non-canonical bespoke one-offs.test-results/,playwright-report/) which brokeformat:checkafter any local e2e run.Validation
/generate-weekcold dry-run by a subagent: clean run, 11/11 pages, all spot-checks pass; doc gaps it reported were folded back into the skill/featurecold dry-run by a subagent: shipped feat: add GET /health endpoint for container healthchecks #90 (GET /health) end-to-end with 137 pytest + 41 e2e green; its findings (gh pr create --head, formatter timing, pre-existing-failure triage) folded back into the skill🤖 Generated with Claude Code