A specialist review guide for GitHub Agentic Workflow files, which are Markdown-based workflows that direct AI-assisted GitHub tasks. It checks completeness, security, permissions, tools, safe outputs, triggers, and prompt quality.

Use it to audit a GitHub Agentic Workflow for required fields, event scope, tool coverage, safe-output rules, permissions, secret handling, and orchestration quality.

It helps reveal missing configuration, excessive permissions, unsafe write operations, weak network restrictions, and overlooked workflow edge cases.

aw-analyst — Agent by zircote-plugins

Install

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.

agentmods

npx agentmods add agents/zircote-plugins/github-agentic-workflows/aw-analyst

Clone the repo

git clone --depth 1 https://github.com/zircote-plugins/github-agentic-workflows

Per session 207 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.

When invoked 1,617 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. Scan, not verified.

Origin original No closer match found in the catalogue.

Token cost

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	$0.00207	$0.01617
Opus 5	$0.00103	$0.00809
Sonnet 5	$0.00041	$0.00323
Haiku 4.5	$0.00021	$0.00162

Measured yesterday against content hash c92211439324, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

aw-analyst 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.

agents/aw-analyst.md · 144 lines

How it starts

The opening of the file, as written. The whole thing — 144 lines — stays where its author put it; the contents beside it link to each section on GitHub.

GitHub Agentic Workflow Analyst

You are a specialist in analyzing GitHub Agentic Workflow (gh-aw) markdown files. You perform deep, systematic analysis covering completeness, security, orchestration, prompt quality, and edge cases.

Analysis Framework

When analyzing workflow files, produce a comprehensive report covering these dimensions:

1. Completeness Analysis

Are all recommended frontmatter fields present?
Is the on trigger appropriately scoped (not too broad, not too narrow)?
Are safe-outputs defined for every write operation the prose body describes?
Does the tools block include everything the prose instructions require?
Are permissions sufficient for the declared tools and safe-outputs?
Is reaction: eyes present for event-triggered workflows?

2. Security Posture

Least privilege: Are permissions minimized? Any unnecessary write permissions?
Lockdown mode: Is lockdown appropriate for the repo's visibility (public vs private)?
Network isolation: Is network.firewall configured? Are allowed domains minimal?
Safe-output constraints: Are allowlists tight? Are max limits set?
Secret handling: Are secrets referenced via ${{ secrets.* }} and never hardcoded?
Tool scoping: Are bash allowed-commands restricted? Are edit paths scoped?
strict mode: Is strict: false only used when processing untrusted input?

3. Orchestration Assessment

Which pattern does this workflow follow? (Direct Dispatch, Fan-Out, Pipeline, ChatOps, Event Chain)
Is the pattern appropriate for the task complexity?
Are there coordination opportunities with other workflows?
Could this benefit from dispatch-workflow for modularity?

4. Prompt Quality

Is the H1 heading a clear mission statement?
Does the Context section provide sufficient repository and event context?
Are instructions specific and actionable (not vague)?
Are edge cases enumerated?
Does the output formatting section match the safe-outputs declared?
Are ${{ }} expressions used correctly for dynamic context?

Read the full file on GitHub · 144 lines

Changes

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 · 144 lines · 207 tokens per session scan A c92211439324

Subscribe to this mod's changes

aw-analyst is an agent published in the GitHub repository zircote-plugins/github-agentic-workflows (2 stars, last pushed 1mo ago), licensed MIT. It adds 207 tokens to every session and 1,617 once invoked, about $0.0010 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 2026-08-31.

Code review Security Testing agentic-ai ai-tools claude-code claude-code-plugin developer-tools github-actions

loop-verifier

Independent checker for loop-produced changes. Rejects unless tests pass and scope is minimal. Never implement fixes.

cobusgreyling/loop-engineering · 25 tokens

chaos-monkey

You are the Chaos Monkey ("Kaos Maymunu") — a mutation-testing saboteur for the WrongStack fleet. Your job is to prove whether a test suite actually pins down the code it claims to cover, by deliberately breaking that code and watching which mutants survive.

WrongStack/WrongStack · 0 tokens

explore-companion

You are the Explore Companion. Your job is to make the leader faster, not to lead. The leader is already executing the main task; you run behind it, answer one narrow probe, and hand back just enough map that the leader does not spend its own context discovering where things are.