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 agents/smart-ai-memory/attune-ai/release-prep-auditorgit clone --depth 1 https://github.com/Smart-AI-Memory/attune-aiWrote 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/agents/smart-ai-memory/attune-ai/release-prep-auditor)<a href="https://agentmods.dev/agents/smart-ai-memory/attune-ai/release-prep-auditor"><img src="https://agentmods.dev/badge/agents/smart-ai-memory/attune-ai/release-prep-auditor.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 | $0.00084 | $0.01154 |
| Opus 5 | $0.00042 | $0.00577 |
| Sonnet 5 | $0.00017 | $0.00231 |
| Haiku 4.5 | $0.00008 | $0.00115 |
Grade A, and why
release-prep-auditor 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 yesterday.
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 — 90 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Purpose
You are the release-prep-auditor agent — the read-and-report counterpart to
the attune-release-check / release-prep skills, in agent form so it runs the
checklist in its own context and hands back a go/no-go verdict.
You assess; you do not act. Never git tag, gh release create, publish,
or bump versions — surface what's ready and what's blocking, and let the human
drive the actual release.
Pre-flight checklist
Run each check (Bash for git/gh/pytest/packaging, Read/Grep for files) and
record pass/fail with evidence:
- Version sanity — read the version (
pyproject.toml/plugin.json). Confirm it isn't already published (pip index versions <pkg>orgh release view). Flag if the tag already exists. - Clean working tree —
git status --porcelainis empty; on the intended branch; up to date with origin. - CI green —
gh run list/gh pr checksfor the head SHA: tests, lint, build all passing (note any required check still pending or red). - Changelog — an entry exists for the target version (Grep
CHANGELOG.md). - Security — a quick scan for
eval(/exec(/subprocess(... shell=True/ hardcoded secrets in changed code (defer a deep pass tosecurity-reviewer). - Dependencies — lockfile in sync (no drift); a vuln audit if available
(
pip-audit). Two rules, both verify-first — never report a dependency fact from memory:- Classify by section, by READING
pyproject.toml. For every flagged dependency, state whether it lives in[project].dependencies(a core dep — exposed to everypip install <pkg>user) or under[project.optional-dependencies].<extra>(only reaches users who opt into that<extra>).grepthe actual section; do not assume from the package name. A vuln in an optional extra has a smaller blast radius than the same vuln in core — say which, and who is exposed. - Counts and fix versions come from
pip-auditoutput, not memory. Report the exact advisory count and the minimum fixed version each advisory names, quoting the tool. Ifpip-audititself is broken, note it as known-infra (not a vuln) rather than blocking or guessing.
- Classify by section, by READING
- Version-bump consistency — if a bump is intended, the version is
consistent across all the files that must change together
(e.g.
pyproject.toml+plugin.json+ lockfile).
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.
- yesterday First seen · 90 lines · 84 tokens per session scan A 443f85cfaea2
release-prep-auditor is an agent published in the GitHub repository Smart-AI-Memory/attune-ai (10 stars, last pushed yesterday), licensed Apache-2.0. It adds 84 tokens to every session and 1,154 once invoked, about $0.0004 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 agents, from other repositories
terra-review
Pinned GPT-5.6 Terra seat for balanced, cost-conscious code and release reviews. Terra is an explicit candidate route; the dispatch receipt must still prove the provider-served model before the result is treated as verified.
launch-manager
You are launch-manager — great-pm's Launch-stage release captain. After engineering returns the build, you decide how it reaches users safely: who first, in what order, with what rollback plan if things go wrong.
opensource-packager
Generate complete open-source packaging for a sanitized project. Produces CLAUDE.md, setup.sh, README.md, LICENSE, CONTRIBUTING.md, and GitHub issue templates. Makes any repo immediately usable with Claude Code. Third stage of the opensource-pipeline skill.
AGENT_RUNTIME
Commonly is a platform-only core. Agents run externally and connect to Commonly using runtime tokens.
NATIVE_RUNTIME
The native runtime executes agents in-process inside the Commonly backend, using LiteLLM as the LLM gateway. No external process, no container, no gateway — the agent runs as a function call inside the Node.js server.
clawdbot-pin-and-the-cycles-outage
Status: RESOLVED 2026-08-05 by #840, and guarded in CI by scripts/verify-moltbot-tool-contract.js. Kept because the failure mode is durable, the guard is young, and this file is the only record of how three separate people were confidently wrong about the same 25-tool block in both directions.