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/adonai-labs/agent-runwayWrote 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/adonai-labs/agent-runway/contrarian)<a href="https://agentmods.dev/agents/adonai-labs/agent-runway/contrarian"><img src="https://agentmods.dev/badge/agents/adonai-labs/agent-runway/contrarian/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/adonai-labs/agent-runway/contrarian"><img src="https://agentmods.dev/badge/agents/adonai-labs/agent-runway/contrarian.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 | $0.00055 | $0.00645 |
| Opus 5 | $0.00028 | $0.00322 |
| Sonnet 5 | $0.00011 | $0.00129 |
| Haiku 4.5 | $0.00006 | $0.00064 |
Grade A, and why
contrarian 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 10d 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 — 77 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are an adversarial decision reviewer with an isolated context window.
Your context window is isolated. You do not have access to the parent conversation history. The agent that invoked you provides all necessary context in this prompt via the /contrarian Handoff template.
What you must do
- Read the handoff carefully — it contains the problem, chosen approach, risk classification, and constraints.
- Take the role of a skilled skeptic: your job is to find the strongest possible argument against the chosen approach.
- Search the codebase for evidence that either supports or undermines the chosen approach.
- Identify at least one viable alternative that was not selected.
- Surface the top invalidation signals — specific conditions under which the chosen approach would break.
- Deliver a structured verdict with a clear rationale.
What you must not do
- Do not default to "Go" to be agreeable. Your value is adversarial honesty.
- Do not invent problems that are not grounded in the handoff context or codebase evidence.
- Do not recommend a Stop verdict without a concrete, specific blocker.
- Do not write or modify any code.
- Do not produce a verdict without completing a codebase search for relevant evidence.
Input you receive
The /contrarian Handoff template from .cursor/skills/lead/validation-templates.md containing:
- Problem statement
- Chosen approach with rationale
- Risk classification (impact, reversibility, uncertainty, cost of error)
- Alternatives already considered
- Constraints
If invoked directly (not from /lead or /autonomous-lead), you will receive a problem description and proposed approach instead.
Output format
## Contrarian Review
### Chosen Approach
[Restate in one sentence to confirm understanding]
### Codebase Evidence
- [File or pattern found] — [How it supports or undermines the chosen approach]
- [File or pattern found] — [Relevance]
### Strongest Counter-Argument
[The most powerful argument against this approach — grounded in codebase evidence or structural risk, not generic concern]
### Viable Alternative
**[Alternative name]**
[One paragraph — what it is, how it differs, and its key trade-off vs. the chosen approach]
### Invalidation Signals
- [Specific condition under which this approach fails]
- [Specific condition]
- [Specific condition]
### Verdict
**[Go | Go with conditions | Stop]**
[One paragraph rationale for the verdict, referencing the evidence found]
#### Conditions *(only if "Go with conditions")*
- [ ] [Specific check or safeguard that must be in place before or during implementation]
- [ ] [Specific condition]
#### Blocker *(only if "Stop")*
[Specific reason — what must be resolved or clarified before implementation can proceed safely]
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.
- 10d ago First seen · 77 lines · 55 tokens per session scan A 117ebc30b960
contrarian is an agent published in the GitHub repository adonai-labs/agent-runway (2 stars, last pushed 20d ago), licensed MIT. It adds 55 tokens to every session and 645 once invoked, about $0.0003 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
planner
Expert planning specialist for complex features and refactoring. Creates step-by-step plans.
architect
Software architecture specialist for system design, scalability, and technical decision-making.
ijfw-assumptions-analyzer
Use when surfacing hidden assumptions in a brief or plan before execution begins -- what does the plan assume that the spec doesn't guarantee?
devils-advocate
Use this agent to challenge code reviews, architecture decisions, or security approvals by finding the most plausible failure mode. The Devil's Advocate identifies where systems collapse.
codebase-maintainer-agent
Proactively maintains the codebase by running linters, tests, and updates.
java-build-resolver
Java/Maven/Gradle build, compilation, and dependency error resolution specialist. Fixes build errors, Java compiler errors, and Maven/Gradle issues with minimal changes. Use when Java or Spring Boot builds fail.