Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add commands/damusix/atomic-claude/watch-cigit clone --depth 1 https://github.com/damusix/atomic-claudeWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/damusix/atomic-claude/watch-ci)<a href="https://agentmods.dev/commands/damusix/atomic-claude/watch-ci"><img src="https://agentmods.dev/badge/commands/damusix/atomic-claude/watch-ci.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00069 | $0.01448 |
| Opus 5 | $0.00034 | $0.00724 |
| Sonnet 5 | $0.00014 | $0.00290 |
| Haiku 4.5 | $0.00007 | $0.00145 |
Grade A, and why
watch-ci scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 6d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
Jenkinsfile → Jenkins (curl against $JENKINS_URL) How it starts
The opening of the file, as written. The whole thing — 139 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Step 1 — Pre-flight
git rev-parse --is-inside-work-tree
If not inside a git repo, stop:
not a git repo. /watch-ci needs a versioned project with a CI provider.
Check signals are available — the subagent uses them to identify the CI provider:
test -f docs/wiki/index.md
If missing, warn (don't stop):
note: no docs/wiki/index.md found. subagent will fall back to file-tree heuristics
(.github/, .gitlab-ci.yml, .circleci/, Jenkinsfile, etc.).
suggest running /refresh-wiki after to improve future runs.
Step 2 — Capture context for the brief
git rev-parse --abbrev-ref HEAD # current branch
git rev-parse HEAD # current SHA
git remote -v | head -1 # remote URL — provider hint
pwd # absolute repo path
$ARGUMENTS may narrow the target. Pass it through verbatim — the subagent classifies it using provider-appropriate semantics:
- empty → current branch's latest run(s)
- integer → run / pipeline ID
#NorN→ PR / MR number- branch name → that branch's runs
- workflow / pipeline filename (
.yml,.yaml) → that specific workflow
Don't classify in the foreground. Different providers use different shapes (GitHub run IDs, GitLab pipeline IDs, CircleCI workflow UUIDs), so the subagent decides.
Step 3 — Dispatch background subagent
Invoke the Agent tool with:
-
subagent_type: "general-purpose"— generic runner backed by Haiku via per-call model override. -
model: haiku -
run_in_background: true -
description: "Watch CI for <branch-or-target>" -
prompt: a self-contained brief containing:Watch CI for this repo and report when it reaches a terminal state. Repo: <absolute pwd> Branch: <current branch> HEAD SHA: <current SHA> Remote: <remote URL> Target hint ($ARGUMENTS): "<verbatim args, or 'current branch' if empty>" Step 1 — identify the CI provider: a. Read docs/wiki/index.md if it exists. The inferrer's schema includes a CI/CD section that names the provider and points at config files. Use it as the primary source of truth. b. If signals are absent or silent on CI, fall back to file-tree detection: .github/workflows/*.yml → GitHub Actions (use `gh`) .gitlab-ci.yml → GitLab CI (use `glab`) .circleci/config.yml → CircleCI (use `circleci`) Jenkinsfile → Jenkins (curl against $JENKINS_URL) .buildkite/pipeline.yml → Buildkite (use `bk`) bitbucket-pipelines.yml → Bitbucket Pipelines (curl REST API) azure-pipelines.yml → Azure Pipelines (use `az pipelines`) c. If no provider is identifiable, bail: print `no CI provider detected.` and stop. Step 2 — verify the CLI is available: `command -v <cli>` and (where applicable) auth check (e.g. `gh auth status`). If missing or unauthed: print `detected <provider>, but <cli> not installed/authed. install: <command>` and stop. Don't try to fall back to a different tool. Step 3 — resolve the target using $ARGUMENTS-equivalent semantics for the provider: - empty / "current branch" → runs/pipelines on <branch> @ <SHA> - integer → specific run/pipeline ID - `#N` / `N` → PR/MR number; resolve to head SHA, then watch runs on that SHA - branch name → runs on that branch - `.yml`/`.yaml` → specific workflow/pipeline file on <branch> Print one line: `watching <provider>: <resolved target>`. Step 4 — poll until terminal state: Cadence 30-60s. Use streaming watch where the provider supports it (`gh run watch <id> --exit-status`, `glab ci view --live <id>`, etc). Cap total wait at 10 minutes. If still running, report current state and stop. READ-ONLY. Do not rerun, cancel, or modify anything. Step 5 — report concisely: - Per workflow/pipeline: outcome (success/failure/cancelled/timed-out). - On failure: failing step/job name + 1-3 line error excerpt from logs. - Note any side effects (e.g. release-please opened a PR, deployment URL). - One paragraph total. The user is in another conversation; respect their context.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 6d ago First seen · 139 lines · 69 tokens per session scan A 228e864ffd13
watch-ci is a command published in the GitHub repository damusix/atomic-claude (84 stars, last pushed yesterday), licensed MIT. It adds 69 tokens to every session and 1,448 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other commands, from other repositories
deploy
Deploy application with pre/post-deploy checks.
land-and-deploy
This command was migrated to a skill in Claude Code 2.1.3. See: examples/skills/land-and-deploy/SKILL.md.
webhook
Configurar triggers externos (GitHub, Slack, Linear, PagerDuty) para automatizar comandos Don Cheli. Usa cuando el usuario dice "webhook", "triggers", "GitHub webhook", "Slack integration", "automation", "automate Don Cheli", "trigger automation", "CI/CD webhook", "external triggers", "webhook config". Configura…
gate
Crear, listar y ejecutar quality gates custom del proyecto.
ship
Persona: You are the Ship Commander -- the last line of defense before code reaches production. Nothing ships without your approval. Reflection Protocol: See agents/reflection-protocol.md for reflection requirements.
devops
You are the DevOps Specialist -- the single authority on version control, CI/CD pipelines, platform operations (GitHub, Azure DevOps, GitLab), infrastructure, deployment, backup, and cleanup. If it touches git, pipelines, or production infrastructure, it's yours.