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 commands/aaronbassett/agent-foundry/adversarialgit clone --depth 1 https://github.com/aaronbassett/agent-foundryWrote 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/commands/aaronbassett/agent-foundry/adversarial)<a href="https://agentmods.dev/commands/aaronbassett/agent-foundry/adversarial"><img src="https://agentmods.dev/badge/commands/aaronbassett/agent-foundry/adversarial.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 | $0.00037 | $0.00678 |
| Opus 5 | $0.00018 | $0.00339 |
| Sonnet 5 | $0.00007 | $0.00136 |
| Haiku 4.5 | $0.00004 | $0.00068 |
Grade A, and why
adversarial 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 — 72 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/decision-making:adversarial
When to use
Use this command when you already have N≥2 concrete options and want a quick tiebreak with steelmanned briefs. It is intended for reversible or moderate-stakes decisions where a single round of advocacy is enough to surface the strongest case for each option.
Escalate to /decision-making:tribunal instead if you want rebuttals,
iteration, multi-round fact-finding, or an honest "unable to decide" escape
hatch. adversarial is explicitly lightweight — one round, no rebuttals, no
judge questions.
Cost tier
Medium. N parallel general-purpose subagents, one round, no iteration. See references/cost-tiers.md.
Input
A decision description plus an explicit list of N≥2 options. If the caller's arguments do not contain a clear decision and at least two options, ask once and only once for them. No multi-round fact-finding — that is tribunal's job. Lightweight means lightweight.
Workflow
-
Parse options — extract N options from the caller's arguments. If they are missing, ask the user once for a decision description and the explicit list of options. If the user still does not provide them, produce an error message recommending
/decision-making:tribunalfor fuzzy decisions that need multi-round fact-finding. -
Spawn advocates in parallel — spawn one
general-purposesubagent per option, all in parallel in a single tool-call batch. Each subagent is given ONLY its assigned option plus the shared decision context — never the competing options' briefs. Use this exact prompt template, substituting the bracketed slots:You represent: [OPTION] Decision context: [WHAT IS BEING DECIDED] Background: [ANY CONTEXT THE USER PROVIDED] Present the strongest possible steelmanned case for why [OPTION] is the best choice. Anticipate the most likely rebuttals and address them pre-emptively. Aim for 200-400 words. Good-faith rules: no fabricated facts, no straw-manning competing options, acknowledge genuine weaknesses honestly, commit fully to [OPTION] without equivocation.
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 · 72 lines · 37 tokens per session scan A 5e108ea549d7
adversarial is a command published in the GitHub repository aaronbassett/agent-foundry (4 stars, last pushed 20d ago), licensed MIT. It adds 37 tokens to every session and 678 once invoked, about $0.0002 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 commands, from other repositories
checklist
Generate a custom checklist for the current feature based on user requirements.
clarify
Identify underspecified areas in the current feature spec by asking up to 5 highly targeted clarification questions and encoding answers back into the spec.
specify
Create or update the feature specification from a natural language feature description.
converge
Assess the current codebase against the feature's spec, plan, and tasks, then append any remaining unbuilt work as new tasks to tasks.md so implement can complete it.
implement
Execute the implementation plan by processing and executing all tasks defined in tasks.md.
analyze
Perform a non-destructive cross-artifact consistency and quality analysis across spec.md, plan.md, and tasks.md after task generation.