Add Claude Code skills for LIF Core (dev loop, PR lifecycle, ops, planning)#992
Open
bjagg wants to merge 4 commits into
Open
Add Claude Code skills for LIF Core (dev loop, PR lifecycle, ops, planning)#992bjagg wants to merge 4 commits into
bjagg wants to merge 4 commits into
Conversation
Port four agent skills from tabletop-platform, re-lensed for LIF Core's stack (Python Polylith + FastAPI + React/TS frontends): - test: full check suite (ruff/ty/pytest/poly + integration + advisor-app vitest) - refactor: atomic-change-then-verify loop with uv-based verify gate - adr-write: knows the per-domain-folder, per-folder-numbered ADR layout - multi-agent-plan: Workflow of sequential Opus Plan agents, 4 lenses (FP/Polylith -> backend correctness -> frontend/holistic) Skipped tabletop's code-review skill (lif-core already has built-in /code-review + /security-review) and session-wrapup (wrapper-level concern). cspell: add "elif" and "threadpool" to the word list. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Three more agent skills for LIF Core operations: - new-microservice: executable form of the adding-a-new-microservice runbook (base + project + auth + compose + CI + CFN + tests), with the trimmed-deps / PEP440 / per-project-pyproject / brick-match pitfalls baked in. - promote-to-demo: dev -> demo promotion (release-demo.sh tag bump -> aws-deploy.sh stacks -> MDR frontend -> SAM-if-needed), gated on dev validation + explicit sign-off; sequential-deploy + reset-DB guards. - issue-sweep: GitHub-issue analogue of a backlog-drift sweep. Workflow fans out cheap judges per open issue (resolved-by-merged-code?), arbitrates disagreements, surfaces an evidence-backed close-list; never auto-closes. cspell: add "oneline" to the word list. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
…ress-pr-feedback) Three skills matching the observed PR workflow in this repo: - self-review-relay: Workflow of 3 parallel review lenses (correctness/ security, robustness/ops, tests/conventions/scope) + a synthesis pass that sorts blockers vs nits. The rigor step before requesting review. - open-pr: scaffolds a convention-compliant PR. Encodes the gotcha that only CLAUDE.md auto-loads -- AGENTS.md + CONTRIBUTING.md + the PR template must be read each time. Fills the template, enforces Issue #XXX commits + green pre-commit + one-issue-per-PR + no-self-approve. - address-pr-feedback: triage reviews + inline comments, fix/clarify/push-back/ defer, append commits (never force-push per feedback_pr_commit_style), reply inline, re-request review. Outward-facing replies gated on confirmation. cspell: add "aget" (LangGraph async method prefix). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Dry-run findings: passing the issue list via the Workflow args field arrives undefined (pipeline() throws). Inline the Phase-1 list as a const instead. Also document the cost (~75K tokens/issue) and default to a label filter or stalest-N rather than sweeping all open issues. Co-Authored-By: Claude Opus 4.8 <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.
Description of Change
Adds 10 Claude Code skills under
.claude/skills/(alongside the existingdocs-indexskill), each encoding a recurring LIF-Core workflow as an executable, convention-aware procedure so they run consistently instead of being re-derived ad hoc each time.pyprojectrule, fail-closed tenant routing, single-uvicorn-worker async, append-don't-force-push, "onlyCLAUDE.mdauto-loads").test(ruff/ty/pytest/poly + integration + vitest),refactor(atomic-change→verify),adr-write(per-domain-folder ADR layout)self-review-relay(multi-lens pre-review pass),open-pr(reads AGENTS.md/CONTRIBUTING.md/template, base-branch + fork checks),address-pr-feedback(triage + reply + re-request)new-microservice(the add-a-service runbook),promote-to-demo(dev→demo gated promotion),multi-agent-plan(sequential design passes),issue-sweep(multi-agent stale-issue audit)cspell.jsonwords. No application code, runtime, or deployment config is touched.multi-agent-planandissue-sweepuse the Workflow tool and explain a fallback when it isn't available./test,/open-pr) in a LIF-Core session;cspellpasses; no code paths are affected.Related Issues
No tracking issue — developer-experience tooling. Happy to file one if the team prefers.
Type of Change
Project Area(s) Affected
.claude/skills/Checklist
uv run ruff check) — N/A to Markdown; pre-commit ran clean on every commitTesting
cspellclean; skill frontmatter/structure verifiedAdditional Notes
code-reviewskill (LIF-Core already has built-in/code-review+/security-review), plussession-wrapupand several domain-specific skills that don't apply here.