Overload governance
ACP is a reusable democratic coordination primitive for discussion under load. Relay is the reference implementation used to validate that primitive in practice.
npx -y skills add jasontang-ai/acp --skill overload-governanceAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Use this skill to enforce ACP digest size limits, bridge budgets, pacing, and bounded participant reading burden.
SKILL.md
2.6 KB, as published. Nobody here has run it
What this skill is for
Use this skill to keep ACP deliberation usable under load by checking digest size, bridge exposure, and reading burden before release.
When to use it
- Reviewing a cycle config before routing or release.
- Checking whether digest item counts exceed configured limits.
- Interpreting overload metrics in benchmark or pilot evidence.
When not to use it
- Do not use it for generic analytics.
- Do not use it to suppress minority views solely because they add complexity.
- Do not use it as a substitute for routing or fairness review.
Inputs expected
- Cycle config
- Digests or routing outputs
- Participant count and contribution count
- Optional benchmark metrics
Preflight / prerequisites
- Confirm configured
maxDigestItemsandmaxBridgeItems. - Confirm whether the cycle is baseline or intervention.
- Inspect the actual digest set, not only aggregate counts.
Workflow
- Count digest items per participant.
- Count bridge items per digest.
- Compare counts to cycle config.
- Look for repeated dominant-view items that consume attention budget.
- Recommend release, revise, or escalate.
Decision rules / judgment criteria
- Hard config limits block release.
- Bridge exposure must be bounded, justified, and not tokenistic.
- Lower reading burden is not acceptable if it erases material issues.
- Load governance must be paired with omission review.
Escalation rules
Escalate when digest budget is exceeded, bridge budget is exceeded, unresolved questions are dropped for space, or load reductions erase a minority signal.
Available scripts
scripts/check-load.mjs: inspect digest item counts and bridge counts against cycle load limits.
Outputs
- Load budget assessment
- Bridge budget assessment
- Overload risk notes
- Release/revise/escalate recommendation
Failure handling
If config is missing or invalid, stop and repair the cycle definition before judging release readiness.
Trust / safety notes
Load limits are constitutional constraints, not UX preferences. They must be explicit and inspectable.
Composition notes
Use with epistemic-routing, digest-and-explanation, and omission-critic in any high-volume workflow.
Examples to inspect next
Read examples/casebook.md for budget pass, budget failure, escalation, and anti-pattern cases.
Evaluation hooks
Run scripts/check-load.mjs and npm run skills:audit.