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 skills/blueberrycongee/termcanvas/challengenpx skills add blueberrycongee/termcanvas --skill challengegit clone --depth 1 https://github.com/blueberrycongee/termcanvasWrote 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/skills/blueberrycongee/termcanvas/challenge)<a href="https://agentmods.dev/skills/blueberrycongee/termcanvas/challenge"><img src="https://agentmods.dev/badge/skills/blueberrycongee/termcanvas/challenge.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 | $0.00041 | $0.01346 |
| Opus 5 | $0.00020 | $0.00673 |
| Sonnet 5 | $0.00008 | $0.00269 |
| Haiku 4.5 | $0.00004 | $0.00135 |
Grade A, and why
challenge 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 4d 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 — 158 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Challenge
Multi-angle adversarial review via isolated Hydra workers. Each worker attacks the same input using a different analytical method, with no visibility into the others' reasoning.
When to use
- User says "challenge this", "stress-test this", "poke holes in this", "what am I missing", "argue against this", or similar
- User has been discussing a topic and wants independent critical review
- Any argument, proposal, opinion, decision, or design that needs pressure-testing
Step 1: Extract
Summarize the argument/proposal/opinion from the current conversation into a neutral, complete brief. Include:
- The core claim or proposal
- Key supporting reasons the user or you have discussed
- Any constraints or context that are relevant
Do NOT editorialize or signal which parts you think are weak. The summary must be fair — biased summaries defeat the purpose.
Step 2: Spawn 4 workers
Use hydra spawn to launch 4 parallel workers. Each worker receives the
same summary but a different methodology prompt. Inherit the current
terminal's provider via --worker-type.
Worker prompts
Each worker prompt must include, in this order:
- The mandatory preamble below (verbatim)
- The full summary from Step 1
- The methodology instructions below (one per worker)
- Instruction to write findings to
result.jsonatomically
Mandatory preamble (prepend to every worker prompt verbatim)
SCOPE RULE — strictly enforced. Your analysis MUST extend beyond the immediate input. The input is your starting point, not your boundary. You are required to:
- Follow every chain. When you find something, do not note it and move on. Ask "what does this lead to?" and trace it at least 2-3 links further. Each link must be a concrete step, not a vague worry.
- Search outward. For every finding, actively look for evidence from outside the input's immediate context — other fields, other systems, historical precedents, known failure cases, research, prior art. If you cannot name a specific external reference, you have not searched wide enough.
- Refuse shallow answers. If a finding can be stated in one sentence with no chain and no external reference, it is not finished. Deepen it or discard it.
A review that stays inside the input's own frame is a failure. You will be evaluated on depth of chains and breadth of external evidence.
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.
- 4d ago First seen · 158 lines · 41 tokens per session scan A 6043921f193d
challenge is a skill published in the GitHub repository blueberrycongee/termcanvas (393 stars, last pushed 3mo ago), licensed MIT. It adds 41 tokens to every session and 1,346 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 2026-08-30.
Other skills, from other repositories
fix-issues
Auto-fix GitHub issues labeled as bugs: fetch open bug issues, analyze feasibility, fix code, and submit PRs. One issue per invocation. Use when: (1) User says "/fix-issues", (2) User asks to fix GitHub issues.
fix-sentry
Auto-fix high-frequency Sentry issues: fetch issues > N occurrences, analyze stack traces, fix code, create GitHub issues, and submit PRs. Supports user feedback issues (event.type "default") with attachment analysis (logs, screenshots) when includefeedback=true. Use when: (1) User says "/fix-sentry", (2) User asks to…
pr-automation
PR Automation Orchestrator: poll open PRs, check CI, run review, fix, and merge eligible PRs. Use when: (1) Invoked by daemon via scripts/pr-automation.sh, (2) User says "/pr-automation".
pr-verify
PR Verification & Merge: verify bot:ready-to-merge PRs with impact analysis, test supplementation, and one-click merge. Use when: (1) User says "/pr-verify", (2) User wants to verify and merge ready PRs.
content-haro-reply
Write a HARO (Help A Reporter Out) or Qwoted/SourceBottle expert-source reply that gets quoted - not buried. Opens with a credentialed one-liner that answers "why this person, for this query," delivers the answer in 3-5 tight bullets with specifics and a contrarian take, and closes with a pull quote written to be…
convert-package
Build a complete sales asset end-to-end via The Donahoe Method: diagnostic → primitives → fingerprint → page → integrity-gate → visual → audit → report. Composite orchestrator that calls 14-16 sub-skills via delegatetask in an 8-phase DAG (sequential where dependencies require it, parallel where they don't). Produces…