agent-build-decide
Skill1 starsChanged 5 months agoApache-2.0View on GitHub
Synthesize one Agentic Build change packet into a short decision surface for Brian and recommend a call. Brian decides — this verb does NOT decide, merge, or edit the packet's work. Use when explicitly asked to run agent-build-decide, or to turn a reviewed packet into Brian's decision surface. NOT the independent critique (that's agent-build-review).
In this repository10 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- agent-build-proposeSkill
- agent-build-reviewSkill
- agent-learning-mineSkill
- commit-allSkill
- learning-reviewSkill
- market-read-labSkill
- review-changeSkill
- work-menuSkill
The file
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 3 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- could be automatic: a setting could enforce it
- Distinctive: line 21
Read before writing — these carry the format and the rules.
- Distinctive: line 30
Read the packet's docs and the context above.
- Distinctive: line 41
Stop. Don't decide, don't merge, don't edit the packet's work.
- Very specific: line 45
Write the surface to decision-surface.md in the packet.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
$ARGUMENTS
Purpose
Turn one change packet's current state into a decision surface for Brian: the short, plain-English digest he decides from. You synthesize and recommend; Brian decides.
This is the lead's compression step. The review files are detailed for the lead; the decision surface is short for Brian. You do not record the decision, merge anything, or edit the packet's work. One packet per run.
Required context
Read before writing — these carry the format and the rules; don't restate them:
2026-06-21-lead-context.md— the Decision Surface section (what the surface includes and how short it is) plus the lead's risk posture and hard lines.- The packet's own docs — whatever exists:
proposal.md, anyproposal-review.md/packet-review.md,frame.md,implementation-notes.md. - For the recommendation vocabulary: the Decision Rules in
2026-06-21-proposal.md.
Task
- Read the packet's docs and the context above.
- Synthesize the packet's current state into a decision surface, in the format the lead-context Decision Surface section defines — no more than it asks for.
- State the decision needed for the packet's current stage, and end with a recommendation in the decision vocabulary (the Decision Rules named in Required Context) — clearly a recommendation; Brian decides.
- Report checks honestly. If no independent review exists in the packet, say so on the surface — don't let an unreviewed packet look ready. Absence isn't proof.
- Out-of-scope thing worth remembering? Log an observation per
learning/AGENTS.md— capture only, never a fix. - Stop. Don't decide, don't merge, don't edit the packet's work.
Output
- Write the surface to
decision-surface.mdin the packet. - Also return it inline in your final response (it's already short).
Final response
End with:
- decision-surface path
- the surface itself
- your recommendation — marked as a recommendation; Brian decides
No decision was made; nothing was merged or edited.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

