Category: distribution
Gap: No action.yml, no Dockerfile, no .pre-commit-hooks.yaml at the repo root. packaging/ contains a single Homebrew formula. Shell completions do exist (src/index.ts:247-293, bash/zsh/fish) and install.sh covers curl-install — so the gap is specifically CI and hook-framework distribution.
Why it matters: coco review --severity 7 and coco lint are CI-shaped features with no CI-shaped delivery: every adopter hand-writes a workflow that installs Node 22, npm-installs git-coco, and plumbs OPENAI_API_KEY. The repo's own .github/workflows/ai-review.yml is evidence the demand exists in-house. A pre-commit hook only exists as coco's own bespoke prepare-commit-msg installer (src/commands/hooks/manageHooks.ts), which the millions of repos on the pre-commit framework cannot use.
Sketch: (1) action.yml — composite action wrapping setup-node + npm i -g git-coco@<ref> + a command input, with severity/json passthrough and outputs for the findings JSON. (2) .pre-commit-hooks.yaml declaring a coco-commit-msg hook. (3) Dockerfile on node:22-alpine with gh and glab preinstalled, published from publish-release.yml alongside npm. (4) A README/wiki "Use coco in CI" section.
Already-there leverage: --json, --quiet, --repo, and --severity exit-code gating already make every command CI-friendly; bin/smokeCli.ts gives a ready smoke target for the container; publish-release.yml + update-homebrew-tap.yml show the release-fanout pattern to copy.
Effort notes / risks: M spread across three independent deliverables — file as three issues under one umbrella. The container needs a decision on whether forge CLIs are baked in (size) or expected from the host.
P1 — significant gaps and differentiating bets
Extracted from a repo audit performed 2026-07 (the audit doc it came from was proposed via an unmerged docs PR).
Category: distribution
Gap: No
action.yml, noDockerfile, no.pre-commit-hooks.yamlat the repo root.packaging/contains a single Homebrew formula. Shell completions do exist (src/index.ts:247-293, bash/zsh/fish) andinstall.shcovers curl-install — so the gap is specifically CI and hook-framework distribution.Why it matters:
coco review --severity 7andcoco lintare CI-shaped features with no CI-shaped delivery: every adopter hand-writes a workflow that installs Node 22, npm-installsgit-coco, and plumbsOPENAI_API_KEY. The repo's own.github/workflows/ai-review.ymlis evidence the demand exists in-house. A pre-commit hook only exists as coco's own bespokeprepare-commit-msginstaller (src/commands/hooks/manageHooks.ts), which the millions of repos on thepre-commitframework cannot use.Sketch: (1)
action.yml— composite action wrappingsetup-node+npm i -g git-coco@<ref>+ acommandinput, withseverity/jsonpassthrough and outputs for the findings JSON. (2).pre-commit-hooks.yamldeclaring acoco-commit-msghook. (3)Dockerfileonnode:22-alpinewithghandglabpreinstalled, published frompublish-release.ymlalongside npm. (4) A README/wiki "Use coco in CI" section.Already-there leverage:
--json,--quiet,--repo, and--severityexit-code gating already make every command CI-friendly;bin/smokeCli.tsgives a ready smoke target for the container;publish-release.yml+update-homebrew-tap.ymlshow the release-fanout pattern to copy.Effort notes / risks: M spread across three independent deliverables — file as three issues under one umbrella. The container needs a decision on whether forge CLIs are baked in (size) or expected from the host.
P1 — significant gaps and differentiating bets
Extracted from a repo audit performed 2026-07 (the audit doc it came from was proposed via an unmerged docs PR).