hashgraph-online/awesome-codex-plugins / plugins/boshu2/agentops/skills/rpi
rpi
VerifiedDrive one accepted change through implementation and checks to done, with one fresh review only where a mistake is costly. Use when: asked to run or orchestrate work through RPI.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 9.1 KB |
| references/boundaries.md | markdown | 3.9 KB |
| references/outer-goal.md | markdown | 1.5 KB |
references/boundaries.md
View on GitHub# Ownership boundaries for the lean RPI core
RPI owns the authorized outcome through implementation, checks, direct repairs
and, on [Validate's triggers](../../validate/SKILL.md#when-a-fresh-judgment-is-worth-it),
one fresh judgment. Plan shapes missing intent and may revise an approach
falsified by evidence within unchanged accepted outcome/scope. Implement edits
and collects facts. Validate independently judges the exact subject and alone
authors semantic `verdict.v2` when persistence is selected. Memory is optional;
its operation references own recall, mining and curation.
## Native authority
BD or the caller's tracker owns work/status/dependencies/handoffs. Git and
repository policy own content/history and delivery. Native runtimes and callers
own aggregate budgets, work selection, queues, claims, stops and subsequent
outcomes. A skill grants no extra Git, tracker, publishing or credential
permission. Existing caller authorization remains usable; do not invent another
approval step merely because a phase changed. When it covers a commit, use
Implement's [shared-checkout commit method](../../implement/references/change-safety.md#an-authorized-commit-in-a-shared-checkout)
and keep others' pre-run edits out of it (the method's isolated-checkout
transfer covers a file that mixes them), naming those files in the report.
Before an authorized push, list every commit it would publish. If any is
neither this run's nor already under review for that branch, push nothing, keep
this run's commits local and report what was held back. A run that stops short
of done pushes nothing new unless the caller's authorization or repository
policy covers pushing that partial state. Keep one authoritative work account,
not a parallel AgentOps ledger.
The runtime derives exact intent/subject identity, complete changed paths,
receipts and observed context identities. Never invent a model/context identity
or transcribe a fictional runtime packet. New requested proof goes to protected
storage outside Git that the caller selects; preserve legacy `.agents/` proof
under owner policy.
Plans, dashboards and reviews count as subject completion only when requested.
## Direct repair and help
Known failures get direct repair. Evidence that disproves an assumption permits
approach revision under unchanged acceptance and scope. Acceptance or authority
expansion needs caller approval; useful source/generated changes already covered
by a scope class do not. Cheap discriminating checks precede expensive judgment.
Reserve finishing capacity; reuse a check result while its inputs are unchanged.
Unknown cause, recurrence, no progress or wrong objective is a causal stall, and
[RPI's stall rule](../SKILL.md#causal-stall-and-bounds) sets its one bounded
helper and the limits nothing renews.
## Optional specialists and adapters
Premortem, council, research, factories and runtime adapters are optional. Risk deepens evidence inspection without mandatory specialist dispatch.
A selected factory remains behind its own coordinator, doctor and supervisor
doors; its reconciler creates and repairs sessions. Concurrent writers need
authorized, disjoint source and regeneration scopes in separate isolation
([Orchestrate](../../orchestrate/SKILL.md#check-the-packets)).
Pass bounded task evidence, not the author's desired verdict. Do not start
another runtime merely because it exists.
## Fresh judgment
[Validate](../../validate/SKILL.md) owns the triggers (linked above) and the
verdict rules: reviewer legs, the PASS, FAIL and NOT_PROVEN order, dissent and
running mutating checks only on a disposable copy, never the judged checkout.
Judge legs read; implementers fix, and a check, not another judgment, confirms
a repair. Validate returns judgment, not a repair or delivery
instruction. RPI completes existing authorized work before reporting, within
real bounds; it reports the subject, strongest evidence and any remaining
acceptance gaps, and persists a machine artifact only for a declared consumer
or caller request.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| receiving-code-reviewobra/superpowersUse when receiving code review feedback, before implementing suggestions, especially if feedback seems unclear or technically questionable - requires technical rigor and verification, not performative agreement or blind implementation | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| requesting-code-reviewobra/superpowersUse when completing tasks, implementing major features, or before merging to verify work meets requirements | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| using-git-worktreesobra/superpowersUse when starting feature work that needs isolation from current workspace or before executing implementation plans - ensures an isolated workspace exists via native tools or git worktree fallback | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| verification-before-completionobra/superpowersUse when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |