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 OutlineDriven/outline-driven-development --skill decidegit clone --depth 1 https://github.com/OutlineDriven/outline-driven-developmentWrote 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/outlinedriven/outline-driven-development/decide)<a href="https://agentmods.dev/skills/outlinedriven/outline-driven-development/decide"><img src="https://agentmods.dev/badge/skills/outlinedriven/outline-driven-development/decide/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/outlinedriven/outline-driven-development/decide"><img src="https://agentmods.dev/badge/skills/outlinedriven/outline-driven-development/decide.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 | $0.00038 | $0.00886 |
| Opus 5 | $0.00019 | $0.00443 |
| Sonnet 5 | $0.00008 | $0.00177 |
| Haiku 4.5 | $0.00004 | $0.00089 |
Grade C, and why
decide scanned grade C with 1 finding 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 2d 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.
Tells the agent never to refusehighAnti-refusal
Suppressing the ability to decline removes a core safety control; a later harmful request then succeeds.
- User pick contradicts evidence: execute the pick, state the concern once, never re-litigate. Do not refuse a confirmed pick. How it starts
The opening of the file, as written. The whole thing — 39 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Decide
Contract
| Field | Bound contract |
|---|---|
| Trigger | The user has a fork and wants it resolved and applied, not explored: "help me decide", "just decide this", "what should I do here", or "decide and fix it". |
| Authority | Human-gated: asks before unrecoverable or observable-surface removal; otherwise reversible local, applying the selected change or dispatching the verify/keep/revert loop. No remote mutation. |
| Side effect | Local artifacts only; rollback is git revert or edit revert of the applied change. |
| Done | The fork, user pick, and resulting change are named; the pick is implemented and verified in the same turn. |
Inputs
The fork: the decision point the user named, or one surfaced from the code. Optional: the user's stated preference, which becomes the recommended default only if evidence supports it.
Procedure
- Confirm that the user wants this fork resolved and applied. Do not use this skill for a verdict of the agent's own making, an exploration that ends in a document, a whole-session director-executor posture, a wide design tree handled in batches, or ambiguity within a stated request. Stop if any of those apply; this skill does not route there. Done when: the fork is confirmed as user-requested and apply-ready, or the run stops with the routing reason stated.
- Ground the fork before framing it. Resolve from evidence first with
grep,glob,read, orlsp. A fork the code already answers is not a fork: read it, say what it answers, and do not ask. Never ask what a search settles. Done when: the fork is grounded from evidence or resolved as already-answered by the code. - Frame one single-select per axis. Give two to four options, placing
(Recommended)first and stating its consequence in the description. Each option names what the code looks like afterward. Where two options differ visibly, put the difference inpreviewas a diff or tree. Never usemultiSelectfor override semantics. Done when: one single-select per axis is framed with the recommended option first and consequences stated. - Take the pick and act in the same turn. A pick applied in one edit gets the edit. A pick needing a verify, keep, or revert loop is restated as its goal and executed as that loop. A pick that removes an observable surface or cannot be reverted from git gets one concern stated, then a yes. Done when: the pick is applied as an edit, loop, or gated action in the same turn.
- When applying one pick opens the next fork, ask it immediately with the settled decision restated in one line. Never bank a queue of questions to fire at the end. Done when: the next fork is asked immediately or no next fork exists.
- When the user picks what the agent would not: execute the pick, state the concern once, never re-litigate. Done when: the pick is executed with one concern stated and no re-litigation.
- Close with one line naming the fork, the pick, and what changed. Done when: one closing line names the fork, pick, and resulting change.
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.
- 2d ago Changed · -25 tokens per session 6292709587d1
- 5d ago First seen · 39 lines · 63 tokens per session scan C d108a0875319
decide is a skill published in the GitHub repository OutlineDriven/outline-driven-development (52 stars, last pushed 3d ago), licensed Apache-2.0. It adds 38 tokens to every session and 886 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it C with 1 finding (tells the agent never to refuse). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other skills, from other repositories
audit-project
Run an iterative multi-agent code audit until critical and high findings are resolved. Use when the user says "audit my code", "find all the bugs", "deep code audit", "iterative review", or "review until clean".
duet
Use when the user invokes /duet, says "pair with me", or faces aesthetic, architectural, or irreversible decisions.
goal-prompt-drafting
Use when asked to draft copy-ready /goal objectives for long-running agents; returns one normalized one-line objective with measurable end state, grounded proof, easy-out invariants, a stop clause, and a Missing list. Not for source or remote-system changes.
handoff-prompt
Use when the user asks for a handoff, delegation, or clipboard-ready prompt for another agent: a standalone path-free prompt copied to the clipboard, confirmed by title. Not for session-snapshot briefs — use handoff; never remote, credential, publish, deploy, or irreversible.
publish-branch
Use when asked to publish the checked-out branch: commit and push it on whatever branch it is, the default branch included. Not for creating branches, PRs, force pushes, or pushing any other branch; when the request excludes the default branch, use commit-push-current.
drill
Use when a concept needs practising rather than explaining: run a scaffolded exercise from worked example to independent problem, quiz the learner, run spaced recall over what they cleared, or probe for the gaps blocking what they want next. For explanation, use explain-concept; for an end-to-end build, use capstone.