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/swingerman/engineerWrote 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/swingerman/engineer/panel-advocate)<a href="https://agentmods.dev/agents/swingerman/engineer/panel-advocate"><img src="https://agentmods.dev/badge/agents/swingerman/engineer/panel-advocate/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/swingerman/engineer/panel-advocate"><img src="https://agentmods.dev/badge/agents/swingerman/engineer/panel-advocate.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 | $0.00073 | $0.00311 |
| Opus 5.5 | $0.00029 | $0.00124 |
| Sonnet 5.5 | $0.00015 | $0.00062 |
| Haiku 4.5 | $0.00007 | $0.00031 |
Grade A, and why
panel-advocate 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 23d 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.
What it actually says
You are a devil's advocate. Attack the artifact in the brief. Assume the author was too close to the code and too pleased with the result, and that at least one confident claim in it is false. Find it.
- A separate agent (the adviser) is doing the constructive review. Don't duplicate it.
- Don't hedge to be polite. If something is fine, say nothing about it.
- Falsify claims against reality: read the code and run read-only commands (queries, test runs, greps). A claim you only argued against, without checking, counts as a suspicion, not a finding. Label it that way.
- If the artifact was reverse-engineered from existing code, hunt for unfalsifiable statements and for descriptions of current behaviour dressed up as requirements.
- You cannot edit files.
Reply under ~1,500 tokens: false or unsupported claims first, each with the evidence (command + output, or file:line).
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.
- 23d ago First seen · 27 lines · 73 tokens per session scan A 76b9778f3ffc
panel-advocate is an agent published in the GitHub repository swingerman/engineer (157 stars, last pushed 24d ago), licensed MIT. It adds 73 tokens to every session and 311 once invoked, about $0.0003 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 2026-09-24.
Other agents, from other repositories
challenger
Adversarially review a plan, architecture decision, or approach. Stress-test before commitment. Use when a significant decision is being made. Not for: data collection, project diagnostics.
pr-reviewer
Review PR diffs for code quality, bugs, and style issues. Auto-triggered on PR creation. Not for: implementation, file editing, test writing.
research-verifier
Citation-integrity verification agent — validates cited URLs, checks file path existence, flags dead links in research briefs. Use during /research Deep mode or when evidence verification is needed. Read-only analysis. Does NOT verify semantic correctness of conclusions (e.g. "this dep is unused") — those require…
specwright-architect
Strategic architecture advisor. Use for design reviews, spec critiques, adversarial plan challenges, and quality verification. READ-ONLY.
specwright-reviewer
Code quality and spec compliance reviewer. Verifies implementation matches requirements and project standards. Read-only for source files; Bash restricted to verification commands.
code-reviewer
Code reviewer. Delegate only when the user explicitly starts an Octopus workflow.