Junhanliu-dev/espalier-engineering / skills/espalier-init
espalier-init
VerifiedAnalyze any existing codebase, discover its patterns and best practices, and generate an Espalier structure (rules, skills, wiki, pipeline) so AI coders produce code matching that project's language and conventions. Greenfield repos take the two-pass Decide-Then-Bind path (/espalier-map charts the conventions, init binds them); boilerplate repos run init normally, then map the product. Wires into Claude Code, Codex, and/or GitHub Copilot (.claude/, .agents/skills/ + .codex/, .github/).
templates/rules/development-process.md
View on GitHub# Development Process
## Branch Strategy
{detected from git or CI config}
## Commit Conventions
{detected from git log}
## CI Checks
{existing checks + what they verify}
## Deploy & Verification
{from DISCOVERY.deploy — mechanism, exact command (or "automatic on merge"),
health check, and which environment a merged change reaches. This is what
pipeline Stage 9 runs/verifies. If DISCOVERY.deploy was null, write exactly:
"No deploy configuration discovered — Stage 9 records SKIPPED: no-deploy-config."}
## Quality Bar
{what must pass before merge — derived from CI + project conventions}
## Maintenance Commits
<!-- ESPALIER MAINTENANCE COMMITS v2 — managed anchor; fixed policy text (not scout-discovered), keep this comment -->
- Every change ships as TWO pull requests: the code PR (every path outside
`espalier/`) and its espalier PR (all of `espalier/**` — records, convention
index, and any doc the Stage 7 sweep refreshed). They are siblings on the
canonical branch, not a stack; the espalier PR's body says which to merge
first. Nothing enforces the order.
- Drift YOUR change caused rides YOUR espalier PR, as its own isolated
`docs(espalier):` commit — the Stage 7 sweep writes it, scoped to what this
change's diff falsified. Drift you did NOT cause stays on the weekly
gardener lane: doctor scans and routine prune refreshes ride ONE weekly
maintenance PR on the canonical branch (`canonical-branch` in
`espalier/.espalier-config`). A prune for your OWN critical/expired flag may
ride your espalier PR as its own isolated `docs:` commit; convention
promotions ride the deciding change's espalier PR. Maintenance commits are
never folded into feature commits — the espalier PR is a separate PR, so
that invariant is now structural rather than a matter of discipline.
- The weekly gardener rota: one rotating dev per cadence interval runs the
maintenance loop (worktree of the canonical branch → `/espalier-doctor` →
`/espalier-prune` over flagged files → re-run the doctor if the prune
cleared everything, so the shared `espalier/.doctor-stamp` says `clean` →
one `docs: weekly espalier maintenance` PR). The rota names who — that is
what kills "someone else surely ran it". A skipped week self-corrects: the
stamp ages out and `doctor_due()` nags everyone again.
- Rule-canon gate: every PR touching `espalier/rules/` gets the rule owner's
review. CODEOWNERS routes it automatically ONLY when "Require review from
Code Owners" branch protection is on — without protection the file is
advisory; turn protection on. Rule owners: before approving a rules PR,
fetch canonical state and check for concurrent open rule PRs — no automated
detector can see two PRs that are both still open.
- Union-merge caveat: GitHub's web conflict editor ignores custom merge
drivers (including `merge=union` on `espalier/.ask-gaps.tsv`) — resolve
union-file conflicts locally, never in the web UI.
- Pipeline commits are atomic: the coder commits one bounded unit per
commit at a clean point, staged by path, message per the Commit
Conventions above (`harness-coder.md` → Commit Discipline); the
pipeline pushes them as made and never squashes — squash-merge is the
pull request's policy.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| writing-skillsobra/superpowersUse when creating new skills, editing existing skills, or verifying skills work before deployment | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| xlsxanthropics/skillsUse this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .xltx, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like "the xlsx in my downloads") — and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files (malformed rows, misplaced headers, junk data) into proper spreadsheets. The deliverable must be a spreadsheet file. Do NOT trigger when the primary deliverable is a Word document, HTML report, standalone Python script, database pipeline, or Google Sheets API integration, even if tabular data is involved. | Stars: 197k | Updated | Claude CodeCodex | Verified | Install |
| shipping-and-launchaddyosmani/agent-skillsPrepares production launches. Use when preparing to deploy to production, or when asking what needs to be in place before shipping. Use when you need a pre-launch checklist, when setting up monitoring, when planning a staged rollout, or when you need a rollback strategy. | Stars: 119k | Updated | Claude CodeCodex | Verified | Install |
| childhoodthedotmack/claude-memBabysit a shipped release after merge, publication, or deployment. Verify published packages, production behavior, installed plugins, and late CI or user reports until the observation window is clean; use babysit for PR readiness before release. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |