📚 Add local-ci skill and scripts/ci memory

Teach agents to verify their changes with ./scripts/ci: module list,
task selection flags, log locations under .ci-logs/, and per-module
workflows (lint-only pass, --fix, --paren-repair, common/ consumer
checks). Register the skill in the skills README, point AGENTS.md at
the new memory, and add the script to the critical-info dev scripts.

AI-assisted-by: omen-alpha
This commit is contained in:
Andrey Antukh 2026-09-09 08:59:35 +00:00
parent 6f4a68100d
commit 8a3540336b
5 changed files with 162 additions and 1 deletions

View File

@ -65,6 +65,7 @@ JSON, REPL access, and so on.
| [`nrepl-eval`](skills/nrepl-eval/SKILL.md) | Run Clojure or ClojureScript code in the live REPL sessions (backend and frontend). | | [`nrepl-eval`](skills/nrepl-eval/SKILL.md) | Run Clojure or ClojureScript code in the live REPL sessions (backend and frontend). |
| [`taiga`](skills/taiga/SKILL.md) | Look up Penpot issues, user stories, and tasks in Taiga. | | [`taiga`](skills/taiga/SKILL.md) | Look up Penpot issues, user stories, and tasks in Taiga. |
| [`testing`](skills/testing/SKILL.md) | The repo's testing rules and TDD workflow, loaded before writing tests. | | [`testing`](skills/testing/SKILL.md) | The repo's testing rules and TDD workflow, loaded before writing tests. |
| [`local-ci`](skills/local-ci/SKILL.md) | Run CI-style lint, test, and format checks for the modules you touched with `scripts/ci`, and read the logs when they fail. |
| [`security-and-hardening`](skills/security-and-hardening/SKILL.md) | Security checks for code that handles user input, auth, or external services. | | [`security-and-hardening`](skills/security-and-hardening/SKILL.md) | Security checks for code that handles user input, auth, or external services. |
| [`ste`](skills/ste/SKILL.md) | Rewrites prose in Simplified Technical English. Loads only when you name it. | | [`ste`](skills/ste/SKILL.md) | Rewrites prose in Simplified Technical English. Loads only when you name it. |
| [`refine-prompt`](skills/refine-prompt/SKILL.md) | Rewrites a rough prompt into a clearer one. Never runs the prompt. | | [`refine-prompt`](skills/refine-prompt/SKILL.md) | Rewrites a rough prompt into a clearer one. Never runs the prompt. |

View File

@ -0,0 +1,95 @@
---
name: local-ci
description: Run local CI-style checks with ./scripts/ci (lint, tests, format) per monorepo module. Use when verifying changes before declaring work done, running lint or tests locally, fixing formatting, or repairing Clojure delimiter errors.
---
# Local CI
Run the same checks CI runs, locally, for the modules you touched, with
`scripts/ci`. Each task writes a log file; the final summary says what
passed and what failed.
Full details: `mem:scripts/ci` (file: `.serena/memories/scripts/ci.md`)
## When to use
- After implementing or fixing code — verify every module you touched
before declaring the work done.
- When the user asks to run CI, lint, tests, or format checks locally.
- When you changed `common/` — validate its consumers too.
**Skip:** while exploring, planning, or reading code.
## Command reference
Run from the repo root:
```bash
./scripts/ci [OPTIONS] [MODULES...]
```
Modules: `frontend` `backend` `common` `render-wasm` `exporter` `mcp`
`plugins` `library`, or `--all` for every module.
With no task flags it runs three tasks per module, in order: **lint**,
**test**, **fmt** (format check; `--fix` formats files instead).
| Flag | Effect |
|------|--------|
| `--all` | Run every module |
| `--exclude MOD` | Skip one module (repeatable) |
| `--lint` / `--no-lint` | Run only lint / drop lint |
| `--test` / `--no-test` | Run only tests / drop tests |
| `--fmt` / `--no-fmt` | Run only format check / drop it |
| `--fix` | Format files instead of checking (other tasks unaffected) |
| `--paren-repair` | Fix delimiter errors in Clojure/CLJS files |
| `--fail-fast` | Stop at the first failure |
| `--quiet` | Suppress failure output |
| `--dry-run` | Show what would run, execute nothing |
| `--clean` | Delete the `.ci-logs/` directory |
## Reading failures
Every task writes its full output to `.ci-logs/<module>-<task>.log`. On
failure the script prints only the last 30 lines. To diagnose a failure,
**read the log file** — never re-run the command piped through filters
(repo rule: redirect to a file first, then read it). The exit code is 1
when any task failed; the summary lists each failed `module:task` and its
log path.
## Typical workflows
```bash
# Verify a module you changed: lint + tests + format check
./scripts/ci frontend
# Fast pass while iterating: lint only
./scripts/ci --lint frontend
# Lint + format check, skip the long test suite
./scripts/ci --no-test frontend
# Format the module without running the test suite
./scripts/ci --fix --no-test frontend
# Broke delimiters in Clojure/CLJS files: repair first, then lint
./scripts/ci --paren-repair frontend
./scripts/ci --lint frontend
# Changed common/ — validate its consumers too
./scripts/ci frontend backend exporter
# Preview what would run, without running it
./scripts/ci --dry-run --all
```
## Gotchas
- Run from the repo root.
- Test tasks are long-running (backend runs `clojure -M:dev:test`); give
the bash call a generous timeout (1020 minutes) instead of letting it
time out mid-run.
- `mcp` has no lint task — it shows as skipped, not failed.
- `--paren-repair` only fixes delimiters; run lint afterwards to catch
what remains. See `mem:scripts/paren-repair`.
- What to run and how to read test results: `mem:testing`.

View File

@ -78,6 +78,10 @@ module. You can read it from `mem:<MODULE>/core`
workspaces (root, modules, member packages). Keeps the shared pnpm store workspaces (root, modules, member packages). Keeps the shared pnpm store
at `<repo>/.pnpm-store` unless `--store`; ignores `external/` and at `<repo>/.pnpm-store` unless `--store`; ignores `external/` and
`.opencode/`. Usage and reinstall steps: `mem:workflow/updating-pnpm`. `.opencode/`. Usage and reinstall steps: `mem:workflow/updating-pnpm`.
- `scripts/ci` — CI orchestration script: runs lint, tests, and format
checks per module (`frontend backend common render-wasm exporter mcp
plugins library`). Logs go to `.ci-logs/`; read the log file on failure.
See `mem:scripts/ci`.
# Dependency graph # Dependency graph

View File

@ -0,0 +1,61 @@
# CI (scripts/ci)
`scripts/ci` runs CI-style checks — lint, tests, format — for one or more
monorepo modules and prints a per-task summary. It is the local equivalent
of CI; use it to verify changes before declaring work done.
## When to use
- After implementing or fixing code in a module: run its checks before
finishing (AGENTS.md: run the applicable lint and format checks).
- When `common/` changed: validate its consumers too (frontend, backend,
exporter; see the dependency graph in `mem:critical-info`).
- To fix formatting across a module (`--fix`) or repair delimiters
(`--paren-repair`) before linting.
## How to use (CLI)
Run from the repo root:
```bash
./scripts/ci MODULE... # lint + test + fmt per module
./scripts/ci --all --no-test # lint + fmt on all modules
./scripts/ci --lint frontend # lint only
./scripts/ci --fix --no-test frontend # format files, skip tests
./scripts/ci --paren-repair --all # fix delimiters in all Clojure modules
./scripts/ci --dry-run --all # preview what would run
```
Modules: `frontend backend common render-wasm exporter mcp plugins library`.
Flags:
- Default tasks: `lint`, `test`, `fmt` (format check; `--fix` formats
instead).
- `--lint` / `--test` / `--fmt` run one task only; `--no-lint` /
`--no-test` / `--no-fmt` drop one task from the default set.
- `--paren-repair` runs only the delimiter repair — it wraps
`scripts/paren-repair` over each module's Clojure/CLJS sources; see
`mem:scripts/paren-repair`.
- `--all` selects every module; `--exclude MOD` drops one (repeatable).
- `--fail-fast` stops at the first failure; `--quiet` suppresses failure
output; `--dry-run` prints commands without running; `--clean` removes
the log directory.
## Logs and exit codes
- Full output of every task: `.ci-logs/<module>-<task>.log`.
- On failure the script prints the last 30 lines; the final summary lists
every failed `module:task` with its log path.
- Exit code 0 when all selected tasks passed, 1 otherwise.
- Diagnose failures by reading the log file — never pipe test output
through filters (AGENTS.md hard rule).
## Notes
- `mcp` has no lint task (shows as skipped). `render-wasm` uses `./lint`,
`./test`, and `cargo fmt`.
- Test tasks are long-running (backend: `clojure -M:dev:test`); use a
generous timeout when calling it from an agent shell.
- Skill entry point: `.agents/skills/local-ci/SKILL.md`.
- Testing principles and output discipline: `mem:testing`.

View File

@ -145,6 +145,6 @@ precision while maintaining a strong focus on maintainability and performance.
- `scripts/nrepl-eval.mjs` — Evaluate Clojure code via nREPL (backend + frontend). - `scripts/nrepl-eval.mjs` — Evaluate Clojure code via nREPL (backend + frontend).
- `scripts/check-commit` — Validate commit messages against Penpot's commit guidelines. - `scripts/check-commit` — Validate commit messages against Penpot's commit guidelines.
- `scripts/check-fmt-clj` — Check Clojure formatting without modifying files. - `scripts/check-fmt-clj` — Check Clojure formatting without modifying files.
- `scripts/ci` — CI orchestration script for running lint, tests, and format checks across modules. See `scripts/ci --help`. - `scripts/ci` — CI orchestration script for running lint, tests, and format checks across modules. See `mem:scripts/ci`.
- `scripts/gh.py` — Multi-purpose GitHub CLI helper. Subcommands: `issues` (list issues in a milestone), `prs` (fetch PR details), `advisories` (list/inspect security advisories). See `python3 scripts/gh.py --help`. - `scripts/gh.py` — Multi-purpose GitHub CLI helper. Subcommands: `issues` (list issues in a milestone), `prs` (fetch PR details), `advisories` (list/inspect security advisories). See `python3 scripts/gh.py --help`.