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 skills add davidomil/cloudx --skill cooperation-feedbackgit clone --depth 1 https://github.com/davidomil/cloudxWrote 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/davidomil/cloudx/cooperation-feedback)<a href="https://agentmods.dev/skills/davidomil/cloudx/cooperation-feedback"><img src="https://agentmods.dev/badge/skills/davidomil/cloudx/cooperation-feedback/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/davidomil/cloudx/cooperation-feedback"><img src="https://agentmods.dev/badge/skills/davidomil/cloudx/cooperation-feedback.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 | $0.00040 | $0.00478 |
| Opus 5.5 | $0.00016 | $0.00191 |
| Sonnet 5.5 | $0.00008 | $0.00096 |
| Haiku 4.5 | $0.00004 | $0.00048 |
Grade A, and why
cooperation-feedback 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.
What it actually says
Answer feedback where it was asked
Draft skill for the proposed helper in protocol.md. Follow cooperation-board throughout the task.
- Read current context and
inbox list. Claim only an event addressed to this tab and assignment generation withinbox claim --event ID. Do not answer every event on a shared board or claim another tab’s work. - Read the target, original quote, block/revision anchor, thread, connected items, and WIP state. If its anchor is stale or deleted, preserve the quote. Explain against that quote or ask for reattachment; do not guess a new target.
- Answer the actual question in its existing thread with a few short bullet points. Put the answer first. Give one fact, action, or question per bullet. Follow the ASD-STE100 principles in cooperation-board. Keep conditions, uncertainty, and source links.
- Submit one transaction that appends your message and names its
respondsToEventIds. Confirm its receipt. This records an answer; only the user resolves the question. Do not delete the question, treat a checkmark as approval, or rewrite the user’s note to hide the concern. - Point briefly to the board thread in the terminal. Leave the viewport, current selection, and other open tools alone.
A question inside WIP authorizes a thread explanation, not a scene edit. Keep comments behind a badge or outside the protected region. Do not expand an obstructive note over the user’s drawing.
For Ready, read the latest region revision and all overlapping WIP regions before acting. An old Ready event cannot release new protection. Change only what the shared task calls for.
Do not start an automatic conversation with another AI’s comments. AI-authored edits are readable context, not invitations to keep waking peers. If the native event was accepted but commit status is unknown, inspect its receipt. If assignment changed, stop using the old board and leave the question there for its next responder. No blind resend or new independent agent session substitutes for delivery to the addressed tab.
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 · 21 lines · 40 tokens per session scan A 4a4a5307bbd8
cooperation-feedback is a skill published in the GitHub repository davidomil/cloudx (1 stars, last pushed 2d ago), licensed MIT. It adds 40 tokens to every session and 478 once invoked, about $0.0002 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 2026-11-03.
cooperation-feedback compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| cooperation-feedbackdavidomil/cloudx | Skill | 1 | 2d ago | MIT | A | 40 | Claude Code, Codex |
| desktop-brand-builderQwenLM/qwen-code | Skill | 29k | yesterday | Apache-2.0 | A | 68 | Claude Code, Codex |
| datavizQwenLM/qwen-code | Skill | 29k | yesterday | Apache-2.0 | A | 26 | Claude Code |
| brand-visual-qualitydevcodex-labs/devcodex | Skill | 432 | 10d ago | AGPL-3.0 | A | 86 | Codex |
| ux-interaction-architecturedevcodex-labs/devcodex | Skill | 432 | 10d ago | AGPL-3.0 | A | 80 | Claude Code, Codex |
| design-system-architecturedevcodex-labs/devcodex | Skill | 432 | 10d ago | AGPL-3.0 | A | 72 | Claude Code, Codex |
| frontend-ux-engineerDominikTobureto/awesome-grok-build | Skill | 69 | 2mo ago | MIT | A | 42 | Claude Code, Codex |