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 skills add flurdy/agent-skills --skill watch-releasegit clone --depth 1 https://github.com/flurdy/agent-skillsWrote 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/skills/flurdy/agent-skills/watch-release)<a href="https://agentmods.dev/skills/flurdy/agent-skills/watch-release"><img src="https://agentmods.dev/badge/skills/flurdy/agent-skills/watch-release/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/flurdy/agent-skills/watch-release"><img src="https://agentmods.dev/badge/skills/flurdy/agent-skills/watch-release.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium Excessive Agency · line 8 Skill selects an external model or provider that may use a different account or billing plan than the operator expects. Undisclosed model switches can cause unexpected cost or quota consumption.Fix: Remove the model/provider override or disclose it prominently and require explicit operator approval before invoking an external coding CLI or billed model.
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 | $0.00054 | $0.02030 |
| Opus 5 | $0.00027 | $0.01015 |
| Sonnet 5 | $0.00011 | $0.00406 |
| Haiku 4.5 | $0.00005 | $0.00203 |
Grade A, and why
watch-release scanned grade A with 0 findings 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 2d 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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 164 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Watch Release
Start a /release-manager loop that runs until 18:00 local time, pacing itself adaptively:
checking again soon (~3 min) while a push is mid-rollout or CI is running, and backing off
(10 → 30 min) once everything is settled. Intended for a dedicated kitty tab during parallel work
— it keeps an eye on git/CI/deploy state and prompts you when something is ready to ship, without
burning a tick every 10 minutes when nothing is happening.
Usage
/watch-release # adaptive cadence, stop at 18:00 (default)
/watch-release 17 # adaptive, stop at 17:00
/watch-release 10m # FIXED 10-minute interval, stop at 18:00
/watch-release 5m 17 # fixed 5m, stop at 17:00
Adaptive is the default. Pass an explicit \d+m interval only when you deliberately want a fixed
cadence (e.g. demoing, or a tab you glance at on a known rhythm).
Instructions
Parse optional arguments:
- Interval — first positive interval matching
\d+m. Absent → adaptive mode (the default). Present → fixed-interval mode at that interval. - Stop hour — first unconsumed stop hour from
0through23. Default:18.
If an argument is invalid, explain the accepted forms and stop. Resolve today's stop hour in local time. If the current time is already at or past that deadline, tell the user and don't start.
Pi protocol v1
If watch_loop is available, use this path instead of Claude scheduling:
-
Call
watch_loopwithaction: status. Continue only when its result reportsprotocolVersion: 1. If another watch isarmed,running, orpaused, do not replace it; show the status and point at/watch-status,/watch-stop, or/watch-resume. If the version differs, explain the mismatch and stop without scheduling anything. -
Convert today's local stop hour to an ISO-8601 timestamp with its timezone offset for
stopAt. -
Use this self-contained prompt in the start call:
Load and follow the skill named `release-manager` now. This is one attended release tick. Invoke the skill rather than improvising its steps, render its full dashboard and every confirmation as visible text, and wait for each answer. Never push without the explicit answer to that action's current question in this tick and release-manager's fresh shared-authority recheck. Never invoke release-maintenance, reconcile manifests, sync config, or restart workloads in a watch tick; an answer cannot expand this boundary. An `ask_user_question` call blocks the active tick until the user answers; do not call `watch_loop` complete while a question is open. After release-manager has finished all answered prompts and printed its final `next-tick:` line, call the matching `watch_loop` action: complete with outcome: continue. In adaptive mode pass that line's numeric N as delaySeconds; if the line is missing or malformed, use 600. In fixed mode omit delaySeconds. Do not complete before the dashboard, answers, tick summary, and cadence recommendation are visible.
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 2d ago Changed · +4 lines · -4 tokens per session 2b46b7deb2da
- 7d ago First seen · 160 lines · 58 tokens per session scan A b31ec34462c6
watch-release is a skill published in the GitHub repository flurdy/agent-skills (9 stars, last pushed 4d ago), licensed MIT. It adds 54 tokens to every session and 2,030 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other skills, from other repositories
chinese-git-workflow
A reference for configuring Git with Chinese code-hosting services such as Gitee, Coding.net, GitLab China, and CNB, including SSH, HTTPS, credentials, CI, and repository mirroring.
baby-sit
Monitor a GitHub pull request until CI is green, diagnose failures, and rerun only evidence-backed flaky GitHub Actions jobs.
atmos-hooks
Atmos hooks: lifecycle events, hook kinds, command/store/git/security hooks, step/steps hooks, when: conditions, scoping and overrides, toolchain integration, --skip-hooks, and Atmos Pro/local output.
atmos-pro
Atmos Pro setup and workflows: settings.pro, GitHub OIDC, affected and inventory uploads, stack locks, pro commit, workflow dispatch, merge queues, and drift detection.
start-temps-cluster
Start (or restart) a local multi-node Temps cluster using Docker-in-Docker — one control plane + 3 worker nodes, each a privileged DinD container running its own dockerd + temps agent, wired with the real multi-host overlay (VXLAN, computecidr allocation) via tools/dev-cluster/ in whichever checkout/worktree you run…
pr-watch
Local PR watcher. Monitors CI status, automatically fixes failing checks by reading failure logs and applying targeted fixes, then optionally merges when all checks pass. Local CLI analog to Claude Code's cloud auto-fix feature.