flag-check
Skill0 starsChanged 3 months agoMITView on GitHub
Use before every step of work to decide whether the team convenes on it or a single owner just does it. Applies the >=50% flag threshold against the active project profile's taxonomy.
In this repository14 other files it gives its agents.
- capability-scanSkill
- conveneSkill
- decision-recordSkill
- deliverySkill
- escalateSkill
- haltSkill
- interface-contractSkill
- merge-guardSkill
- profile-bootstrapSkill
- roster-preflightSkill
- setback-checkSkill
- setback-logSkill
- stability-checkpointSkill
- work-orderSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
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
- Very specific: line 44
Read the active .agentic/profile.md and apply its taxonomy.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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
Flag check
Deciding whether to deliberate is the single biggest cost lever in this system. Convening on everything costs 6-10x a single-model run and buys nothing on work nobody disagrees about. Convening on nothing turns an expensive org chart into one model with extra steps.
The rule
A step convenes when >=50% of the deciding team flags it.
| Team | Members | Threshold |
| Scrum Master | 3 | 2 of 3 |
| UX | 2 | 1 of 2 |
| Dev | 2 | 1 of 2 |
| QA/Sec | 2 | 1 of 2 |
At two-member teams the threshold is unilateral by arithmetic. That is intentional — the threshold exists to stop one cautious voice from convening the three-member council on everything, which is where the cost actually lives.
Standing triggers (every project)
Flag the step if it is:
- Irreversible or expensive to reverse — data migration, deletion, anything a user sees and cannot un-see, anything that changes stored state's meaning.
- Crossing a team interface — the contract between UX and dev, between services, or any public API shape.
- Reversing a prior decision — check
.agentic/DECISIONS.md. Re-litigating a recorded decision always convenes, so the reversal is recorded too. - Ambiguous in the work order — if two readings lead to materially different work, that is a flag, not a judgment call.
- Security, privacy, or money.
Project triggers
Read the active .agentic/profile.md and apply its taxonomy. Those
triggers name what this project gets expensively wrong. If the profile is
missing, run profile-bootstrap — do not fall back to the standing triggers
alone and call it covered.
What does NOT flag
Schema migrations with an agreed shape. Endpoint implementation against an agreed contract. Component styling inside an approved design. Test authoring. Documentation. Anything where the decision was already made and recorded — those have one owner and an async review.
Instrumentation
Record the flag decision (flagged / not flagged, who flagged, which trigger) in the step's Decision Record or work order. Without this, nobody can tell whether the system is flagging everything or nothing — and neither failure is visible in the output, only in the bill. Report the running flag rate to the PM at each gate.
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.