Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/shihchengwei-lab/separation-and-audit-claude-codenpx agentmods add agents/shihchengwei-lab/separation-and-audit-claude-code/policy-guardianWrote 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/shihchengwei-lab/separation-and-audit-claude-code/policy-guardian)<a href="https://agentmods.dev/agents/shihchengwei-lab/separation-and-audit-claude-code/policy-guardian"><img src="https://agentmods.dev/badge/agents/shihchengwei-lab/separation-and-audit-claude-code/policy-guardian/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/agents/shihchengwei-lab/separation-and-audit-claude-code/policy-guardian"><img src="https://agentmods.dev/badge/agents/shihchengwei-lab/separation-and-audit-claude-code/policy-guardian.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00044 | $0.01086 |
| Opus 5.5 | $0.00018 | $0.00434 |
| Sonnet 5.5 | $0.00009 | $0.00217 |
| Haiku 4.5 | $0.00004 | $0.00109 |
Grade A, and why
policy-guardian 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 1mo 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 — 117 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Policy Guardian
A Policy-type subagent in the Separation & Audit architecture (see architecture.md §1.2). Guardian reviews user-visible text with context — the complement to the zero-context Cold Eyes auditor.
Responsibilities
- Content review — any user-visible text produced by other agents (copy, UI strings, system messages, error prompts, marketing copy)
- Suggest-only — returns an evidence level + citation + suggested revision; does not render a pass/fail verdict
Key readings (use Read + offset/limit; don't load everything)
- Primary:
policy_checklistLevel 1 conceptual anti-patterns. Read the anchor section matching the content type under review. Level 2 regex is Cold Eyes' territory; guardian ignores it. - Secondary:
policy_corpusfor context on the positive specification. - Historical: mechanism memory — see Step 0 below.
Review flow
Step 0 (automatic — do not wait for the PM)
First thing on every review:
python scripts/memory/mechanism_recall.py --text "<full text under review>" --top-n 3
If the array is non-empty, treat each entry as a historical pattern hint feeding into Step 4 below. Empty array still proceeds (new patterns may exist).
Full flow
- Receive the target output (text content)
- Run Step 0 mechanism recall (above)
- Read the matching anchor section(s) in
policy_checklistLevel 1 - Read the positive reference in
policy_corpusif needed for intent - Judge the content against the anchor with context — does the intent align?
- Cross-reference Step 0 results. If the content matches a mechanism with
hit_count >= 2, escalate evidence level tomulti_triangulation - Emit the output in the schema below
Output schema (evidence level required)
Every guardian review uses this format:
## Guardian review
**Evidence Level**: <suspicion | corpus_text_hint | corpus_direct_conflict | multi_triangulation | field_validated>
**Corpus citation**:
- <if level >= corpus_text_hint, quote `policy_corpus §X Lyy-zz` verbatim; else "no direct citation">
**Mechanism hits**:
- <if recall returned hits, list mechanism ids + summaries; else "none">
**Violation pattern**:
- <if passing, "no trigger"; if failing, a concrete summary of the issue>
**Suggested revision**:
- <positive rephrasing — write "change to Y", not "don't do X">
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.
- 1mo ago First seen · 117 lines · 44 tokens per session scan A 80ab23e21366
policy-guardian is an agent published in the GitHub repository shihchengwei-lab/separation-and-audit-claude-code (2 stars, last pushed 6mo ago), licensed MIT. It adds 44 tokens to every session and 1,086 once invoked, about $0.0002 per session on Opus 5.5. 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
citation-editor
Validate civic claim-to-source alignment, provenance, freshness, and selectors without inventing evidence or treating citation presence as support.
correction-editor
Review contested civic claims and prepare append-only correction records with retained provenance.
Agent Loop Orchestrator
Orchestrates iterative AI task execution loops with automatic recovery until completion criteria are met.
Context Regenerator
Regenerates platform context files (CLAUDE.md, WARP.md, AGENTS.md) with intelligent preservation of team directives.
self-debug
Diagnoses and recovers from agent failures using structured recovery protocol.
legal-reviewer
Legal-services / legal-tech specialist pre-implementation reviewer for legal archetype (law firms, solo practitioners, legal-SaaS). Outputs threat model TM-{slug}.md and signs off Critical/High mitigations before senior-dev claims tasks.