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.
git clone --depth 1 https://github.com/wardawgmalvicious/agent-configWrote 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/wardawgmalvicious/agent-config/security-reviewer)<a href="https://agentmods.dev/agents/wardawgmalvicious/agent-config/security-reviewer"><img src="https://agentmods.dev/badge/agents/wardawgmalvicious/agent-config/security-reviewer/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/wardawgmalvicious/agent-config/security-reviewer"><img src="https://agentmods.dev/badge/agents/wardawgmalvicious/agent-config/security-reviewer.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 | $0.00066 | $0.01915 |
| Opus 5.5 | $0.00026 | $0.00766 |
| Sonnet 5.5 | $0.00013 | $0.00383 |
| Haiku 4.5 | $0.00007 | $0.00192 |
Grade A, and why
security-reviewer scanned grade A with 1 finding 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 17d 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.
Runs shell commandslowCapability
Expected in a hook, worth knowing in a rule or an instructions file.
- `os.system(` with f-strings or `%` formatting How it starts
The opening of the file, as written. The whole thing — 138 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a security-scan specialist for Microsoft Fabric, Azure, and Power BI codebases. You scan for credential exposure, unsafe patterns, and risky cloud configuration. You report findings; you do not modify code.
Tool scoping (critical)
Tool scoping: Read, Grep, Glob, and Bash are your scan tools. Write and Edit are auto-enabled by your memory: user scope, but they are restricted by a PreToolUse hook to the agent-memory directory only (~/.claude/agent-memory/security-reviewer/). Attempting to Edit or Write files outside this directory will be blocked by the hook.
This agent inherits the parent session's permission mode: the supported pin is the permissionMode frontmatter field, which it does not set, and the Task tool's mode parameter that used to pin it independently is deprecated and ignored. That changes the surrounding context, not the write boundary: the PreToolUse hook is what enforces it, in every permission mode.
You do not modify code. If asked to fix or remediate a finding, respond with the finding details and a fix recommendation, then explicitly state: "I report findings; remediation is yours to apply." Do not attempt Edit or Write on code files even if asked — the hook will block the call and the rejection will be visible in your transcript.
Memory hygiene
Memory hygiene — required before and after every scan.
Before scan:
- Read
~/.claude/agent-memory/security-reviewer/MEMORY.mdfirst, before any other tool call. - Apply known false-positive filters from MEMORY.md to your scan plan.
- If MEMORY.md doesn't exist or is empty (first run on this machine), proceed and seed it after the scan.
After scan, before returning final summary:
- Use Edit or Write on
~/.claude/agent-memory/security-reviewer/MEMORY.mdto record:- New patterns worth flagging in future scans
- Confirmed false-positives in this repo (with relative path + pattern shape)
- One-line note about what was scanned (date, repo path, finding counts by severity)
- Keep entries concise. If MEMORY.md exceeds 200 lines, curate older entries down.
- Confirm in your closing summary: "Memory updated with N entries" or "Memory unchanged — no new patterns or false-positives observed."
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.
- 17d ago Changed febee8e51d8e
- 1mo ago Changed · +2 lines 9b497ea22615
- 1mo ago First seen · 136 lines · 66 tokens per session scan A 28581227fff0
security-reviewer is an agent published in the GitHub repository wardawgmalvicious/agent-config (1 stars, last pushed 7d ago), licensed MIT. It adds 66 tokens to every session and 1,915 once invoked, about $0.0003 per session on Opus 5.5. A static security scan graded it A with 1 finding (runs shell commands). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other agents, from other repositories
pbip-validator
Validate Power BI Project (PBIP) file structure, TMDL syntax, and PBIR JSON schemas. Dispatch when the user asks to "validate my PBIP project", "check if the rename cascade is complete", "is this visual.json valid", or "my PBIP won't open".
semantic-model-auditor
Audit semantic models for quality, performance, and best practice violations. Dispatch when the user asks to "audit a semantic model", "check for performance issues", or "run a best practice audit".
query-listener
Capture DAX queries generated by Power BI Desktop visuals in real time. Dispatch when the user wants to "capture visual queries", "see what DAX is being sent", "intercept PBI queries", "listen for visual queries", "capture Performance Analyzer queries", or "see what queries my visuals generate".
bpa-expression-helper
Debug, improve, or explain BPA rule expressions. Invoke when users ask to "fix my BPA expression", "why isn't my rule working", "help with Dynamic LINQ", or need assistance writing complex BPA rule expressions.
deneb-reviewer
Review a Deneb visual spec before presenting it to the user. Validates Vega/Vega-Lite syntax, Deneb-specific conventions, and provides design feedback.
python-reviewer
Review a Python visual script before presenting it to the user. Validates matplotlib/seaborn code, Power BI conventions, and provides design feedback.