Borrowing it
Nothing to install: this file belongs to ferrus-dev/ferrus. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/ferrus-dev/ferrus/main/.agents/skills/ferrus-supervisor/SKILL.mdgit clone --depth 1 https://github.com/ferrus-dev/ferrusWrote 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/ferrus-dev/ferrus/ferrus-supervisor)<a href="https://agentmods.dev/skills/ferrus-dev/ferrus/ferrus-supervisor"><img src="https://agentmods.dev/badge/skills/ferrus-dev/ferrus/ferrus-supervisor/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/skills/ferrus-dev/ferrus/ferrus-supervisor"><img src="https://agentmods.dev/badge/skills/ferrus-dev/ferrus/ferrus-supervisor.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 | $0.00021 | $0.00561 |
| Opus 5.5 | $0.00008 | $0.00224 |
| Sonnet 5.5 | $0.00004 | $0.00112 |
| Haiku 4.5 | $0.00002 | $0.00056 |
Grade A, and why
ferrus-supervisor 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 12d 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.
Supervisor Operating Playbook
This file is advisory only. Runtime workflow is defined by the initial prompt and exposed Ferrus tools.
Task drafting
- Define the expected outcome clearly
- State relevant constraints and acceptance criteria
- Keep task scope explicit and bounded
- Draft task text that the user can review directly
Review quality
- Judge correctness against the task, not personal preference
- Focus on regressions, missing requirements, and verification gaps
- Write rejection feedback that is concrete and actionable
Consultation quality
- Answer the Executor's actual blocker
- Prefer concrete direction over abstract discussion
- Clarify tradeoffs when there is no single obvious answer
Spec closure quality
- Treat a spec
## Outcomesection as compact project memory - When asked to archive a spec, summarize what actually shipped, deviations, validation evidence, follow-up work, and future context
- Archive only through the exposed
archive_specoperation; the user must approve the exact Outcome before it is saved
Human interaction
- External MCP sessions obtain explicit approval before calling an authoring tool
- Native Nano authoring tools accept ready drafts as proposals; the host confirms once before writing
- In Nano planning, inspect and discuss only; task/spec/archive profiles expose their own authoring tool
- Use
/ask_humanonly when the answer cannot be reliably derived from the repository or current context
Useful Ferrus tools
/enqueue_task(task-definition sessions)/create_spec/archive_spec(HQ archive sessions)/wait_for_review/review_pending/approve/reject/wait_for_consultation/respond_consult/ask_human/wait_for_answer/heartbeat(task-scoped sessions, while owning the task lease)
/create_task is a compatibility tool on unfiltered servers, not role-scoped Supervisor sessions.
Use only tools exposed by the current mode; a Consultant attaches without taking the Executor's lease.
Nano hosts bind reviews/consultations and handle waits/heartbeats. Read review_pending or
consultation_context before deciding; review uses the frozen submission, consultation the
attached Executor workspace. Stop after a confirmed decision. Native authoring uses
definition_context and, for archive evidence, read_archive_artifact.
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 12d ago Changed · +6 lines 1e40914fffc5
- 2mo ago Changed · +6 lines 4f477d4799b7
- 2mo ago First seen · 60 lines · 21 tokens per session scan A b05870cd873a
ferrus-supervisor is a skill published in the GitHub repository ferrus-dev/ferrus (2 stars, last pushed 7d ago), licensed Apache-2.0. It adds 21 tokens to every session and 561 once invoked, about $0.0001 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 2026-08-31.
ferrus-supervisor compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| ferrus-supervisorferrus-dev/ferrus | Skill | 2 | 7d ago | Apache-2.0 | A | 21 | Claude Code, Codex |
| paperclippaperclipai/paperclip | Skill | 111k | yesterday | MIT | A | 54 | Claude Code, Codex |
| index-refreshpaperclipai/paperclip | Skill | 111k | yesterday | MIT | A | 43 | Claude Code, Codex |
| agentmailpaperclipai/paperclip | Skill | 111k | yesterday | MIT | A | 30 | Claude Code, Codex |
| ppt-reviewOrkas-AI/Orkas | Skill | 9.0k | yesterday | MIT | A | 2 | Claude Code, Codex |
| cw-gatesHmbown/Codewhale | Skill | 41k | 1mo ago | MIT | A | 48 | Claude Code, Codex |
| feishuHmbown/CodeWhale | Skill | 41k | 5d ago | MIT | A | 33 | Claude Code, Codex |