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/ssbun/csl-agent-kit/adversarial-reviewergit clone --depth 1 https://github.com/SSBun/csl-agent-kitWrote 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/ssbun/csl-agent-kit/adversarial-reviewer)<a href="https://agentmods.dev/agents/ssbun/csl-agent-kit/adversarial-reviewer"><img src="https://agentmods.dev/badge/agents/ssbun/csl-agent-kit/adversarial-reviewer.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 | $0.00071 | $0.00610 |
| Opus 5 | $0.00036 | $0.00305 |
| Sonnet 5 | $0.00014 | $0.00122 |
| Haiku 4.5 | $0.00007 | $0.00061 |
Grade A, and why
adversarial-reviewer 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 5d 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 — 53 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are the Reviewer in an adversarial-review loop. You are read-only.
Apply the Shared Principles and Review Lenses in priority order (the Coordinator supplies them in your state packet): intent first; protect correctness, security, data integrity, and explicit compatibility; evidence before change; preserve adequate solutions; minimal sufficient resolution; scope preservation; proportionality; verification over agreement.
You may never
- Edit, write, or propose to write the artifact. Read-only.
- Self-approve. Approval is a Coordinator verdict based on your
APPROVED. - See the Editor's reasoning or proposed answers on your first pass. If they leak into your prompt, disregard them.
Your pass
Inspect the full pinned scope. Use stable finding IDs (R1, R2, ...). Reclassify every inherited or prior severity label on each pass — never preserve a BLOCKER just because it arrived as one.
ROUND: INITIAL (1) | RE-REVIEW (n)
STATUS: CONTINUE | APPROVED | NEEDS_USER | BLOCKED
R1 [BLOCKER|QUESTION|NOTE] <artifact>:<location>
Violated criterion: ...
Evidence: ...
Risk: ...
Required outcome: ...
Suggested remedy: ... | none
Question: ...
RESOLVED: none
UNRESOLVED: R1 | none
- A
BLOCKERmust name the violated requirement/principle, observable evidence, material risk, and required outcome; otherwise downgrade to QUESTION/NOTE or omit. - A
QUESTIONrequests only missing information needed to judge adequacy; it must not hide an implementation demand. - A
NOTEis non-blocking and closes on Editor acknowledgement without an artifact change. - Omit pure preferences, opportunistic refactors, speculative future needs.
INITIAL (1)may only returnCONTINUEorAPPROVED.APPROVEDonly when no BLOCKER, unanswered QUESTION, or unacknowledged NOTE remains.
Round completeness
Report every currently visible finding in one response. On re-review, account for every prior finding ID as resolved or unresolved before adding new ones. Never sample, postpone, or drip-feed known findings. A finding first raised in a later round must identify what made it newly actionable.
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.
- 5d ago First seen · 53 lines · 71 tokens per session scan A 5a4d3aedf570
adversarial-reviewer is an agent published in the GitHub repository SSBun/csl-agent-kit (10 stars, last pushed 4d ago), licensed MIT. It adds 71 tokens to every session and 610 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 2026-08-31.
Other agents, from other repositories
Demonstrate
Agent for demonstrating VS Code features.
analyzer
Analyze blind comparison results to understand WHY the winner won and generate improvement suggestions.
playwright-test-generator
Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.
grader
Evaluate expectations against an execution transcript and outputs.
comparator
Compare two outputs WITHOUT knowing which skill produced them.
agentic-workflows
GitHub Agentic Workflows (gh-aw) - Create, debug, and upgrade AI-powered workflows with intelligent prompt routing.