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/skills/espalier-coding.md
View on GitHub---
name: espalier-coding
description: >-
How {project} implements features — the per-layer spec map, the
project-specific implementation checklist, and the lean solution-selection
rule. Load BEFORE writing any code in {project}; used by harness-coder at
Stage 3 (implementation — including its folded test-writing duty), the
contract phase, and on fix rounds.
Triggers: "implement", "add feature", "write code", "which layer does this
go in", "layer spec".
---
# Coding Skill
## How This Project Adds Features
{describe the actual pattern for adding new functionality}
{e.g., "Create a route file, add handler, create service, write repository query"}
## Layer Specs
{for each discovered layer, reference its spec file}
- Read `espalier/skills/espalier-coding/specs/{layer-a}.md` when working in {layer A}
- Read `espalier/skills/espalier-coding/specs/{layer-b}.md` when working in {layer B}
## Implementation Checklist
- [ ] {project-specific step 1}
- [ ] {project-specific step 2}
- [ ] ...
{e.g., "- [ ] Route registered in the route index", "- [ ] Service returns the
project's Result type — never throws across the layer boundary", "- [ ] New
list query is bounded (limit / pagination)"}
## Before Writing Code
The canonical pre-coding sequence (identify layer → read spec → find 1-2
reference files → follow the template exactly → climb the Solution Selection
Ladder) lives in `espalier/agents/harness-coder.md` ("Before Writing ANY
Code") — follow it from there; it is deliberately not restated here so the two
files cannot drift. This skill adds the project-specific parts: the Layer
Specs map and the Implementation Checklist above. Verification is one
call — `exit_gate` with your test files ("Verify in One Call" in the same
agent file) — never a runner, a type-checker, and a formatter in three.
## How This Skill Applies by Stage
The layer map above is stage-agnostic; which parts carry the weight changes
with the context the pipeline loaded this skill in:
- **Stage 3 — implementation (code AND, folded, its tests).** The whole
skill applies: the layer spec for
every layer touched, the full Implementation Checklist, the Solution
Selection Ladder before choosing the change's shape.
- **Test writing (Stage 3 folded duty / contract phase / serial test
pass).** Layer specs still govern WHERE tests live and
what naming/structure they follow; the ladder applies to test code too —
reuse the project's existing test helpers, fixtures, and factories before
writing new ones, and a new test-only dependency is still a NEW dependency
(needs its `requirements.md` line). The test recipes themselves (abuse
tests, failure-mode tests) are canonical in
`espalier/agents/harness-coder.md` and
`espalier/skills/espalier-security/SKILL.md`, not here.
- **Fix rounds — review re-spawns.** Scope is the findings, nothing else:
re-read only the spec(s) for the layers the fix touches, make the smallest
convention-compliant change that clears the finding, and never "improve"
adjacent code the finding didn't name. The checklist applies to what the
fix DID touch; record the delta in coding-report.md.
## Solution Selection (keep it lean)
Best convention-compliant solution wins: conventions first, correctness within
them, clarity then brevity break ties. Reuse what the project already has
(reference files, `espalier/wiki/`) before writing new code; where conventions
are silent on the mechanism, prefer stdlib → native platform feature →
already-installed dependency; NEVER add a new dependency without a
`requirements.md` line naming it. Build nothing the requirement didn't ask
for. The full ladder lives in `espalier/agents/harness-coder.md` — it is the
coder's rule, and the reviewer checks against it (advisory, plus the
new-dependency and cryptic-public-name gates). Write it readable as you go:
named constants over magic values, guard clauses over nesting, small
intent-named functions — defaults canonical in
`espalier/agents/harness-coder.md` ("Write It Readable") and
`espalier/rules/coding-standards.md` ("Readable by Default").
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: 198k | 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 |