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 quanghoangf/vibedoc --skill work-epicgit clone --depth 1 https://github.com/quanghoangf/vibedocWrote 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/quanghoangf/vibedoc/work-epic)<a href="https://agentmods.dev/skills/quanghoangf/vibedoc/work-epic"><img src="https://agentmods.dev/badge/skills/quanghoangf/vibedoc/work-epic/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/quanghoangf/vibedoc/work-epic"><img src="https://agentmods.dev/badge/skills/quanghoangf/vibedoc/work-epic.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 | $0.00131 | $0.01080 |
| Opus 5.5 | $0.00052 | $0.00432 |
| Sonnet 5.5 | $0.00026 | $0.00216 |
| Haiku 4.5 | $0.00013 | $0.00108 |
Grade A, and why
work-epic 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 yesterday.
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 — 49 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Work an epic
Run one epic's tasks in order, one at a time, until VibeDoc says the epic is finished or a human is needed. VibeDoc picks the order: vibedoc_next_task respects Depends on, skips blocked tasks and tasks another agent has claimed, and moves the claimed task to in-progress so the board shows who is working on what. Don't choose tasks yourself, and don't work on a task you did not claim.
Argument: an epic id such as R037. If there is none, ask which epic to work on (vibedoc_get_roadmap lists them). Work on one epic only.
Before the loop
Read the project's CLAUDE.md once. Task files quote the relevant rules, but the commands (build, lint, package manager), bans and commit rules are only in CLAUDE.md. You need them for every task.
The loop
-
Claim. Call
vibedoc_next_task { "epic": "<id>" }. The reply starts with one of three markers:🔨 Claimed— go to step 2. The reply contains the full task file.⏳ Nothing ready— see the stop rules below.✅ Epic … is finished— see the stop rules below.
If a reply ends with a
🗺️ Roadmap out of synchint for this epic (for example, set it to in-progress), apply it withvibedoc_update_roadmap_item, so the roadmap matches the board. -
Read the task. Read its Goal, Scope, Files, Implementation notes, Acceptance criteria and Verify. The task file is the spec. Stay inside its Scope: later tasks in the epic own the rest, and if you do their work now, their diffs will conflict with yours.
-
Implement the task, following the project's patterns.
-
Verify. Run every command in the task's Verify block, and check each acceptance criterion you can check. If something fails and the fix is inside the task's scope, fix it and run Verify again.
-
Done. Only when Verify passes: call
vibedoc_update_task { "taskId": "<id>", "status": "done" }, then commit that task's changes, following the repo's commit rules (message format, attribution, which files to stage). Each task leaves the app in a working state. When you commit after each task, it is safe to stop at any point, and a human can review or revert one task at a time. -
Repeat from step 1. Mark the current task done before you call
vibedoc_next_taskagain. If you don't, the task stays in-progress and the queue treats it as claimed by another agent.
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.
- yesterday First seen · 49 lines · 131 tokens per session scan A aa1d23e19b17
work-epic is a skill published in the GitHub repository quanghoangf/vibedoc (5 stars, last pushed today), licensed MIT. It adds 131 tokens to every session and 1,080 once invoked, about $0.0005 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 2026-09-29.
Other skills, from other repositories
gsd-quick-batch
Batch several /gsd-quick-shaped tasks together — planned, dispatched, and merged as one run.
gsd-pause-work
Create context handoff when pausing work mid-phase.
mcore-split-pr
Split a PR into multiple PRs to reduce the number of required CODEOWNERS reviewer groups.
gait-session-tracking
GAIT session lifecycle management - branch creation, turn recording, audit logging for every NetClaw operation. Use when starting a new NetClaw session, recording a health check or config change, pinning a pre-change baseline, or viewing the audit trail for a troubleshooting session.
close-task-commit-push-pr
Close the active backlog task (detected from branch name), commit all changes, push to remote, and open a pull request. Use when the user says "close task and ship it", "close task commit push pr", or invokes /close-task-commit-push-pr.
omh-ralph-task
Execute one omh-ralph task: file-scope, commit, report.