Goals
July 16, 2026 ยท View on GitHub
Product outcome
AgentOps makes one coding-agent experiment independently inspectable without taking over the consumer's engineering system.
The canonical outcome is:
explicit behavior
-> bounded implementation experiment
-> exact content identity
-> fresh independent judgment
-> durable PASS | FAIL | NOT_PROVEN
Fitness properties
- Behavior before activity. The caller-owned intent states the active behavior, acceptance examples where useful, non-goals, evidence, and bounded write scope before implementation begins.
- One experiment. Implement performs one RED -> GREEN -> refactor cycle and reports facts without retry or delivery authority.
- Fresh judgment. PASS requires explicit, distinct author and validator context IDs plus a freshness attestation.
- Exact subject. Content identity covers files, symlinks, deletions,
executable bits, roots, exclusions, and canonical digests without requiring
Git or
ao. - Honest uncertainty. Mutation, incomplete changed-path coverage, missing identity, or missing proof returns NOT_PROVEN. Proven scope or acceptance failure returns FAIL.
- Sovereign proof. Validate atomically writes content-addressed JSON to caller-controlled storage. Provenance is optional audit.
- Stop boundary. RPI dispatches Plan, Implement, and Validate at most once, reports the result, and emits no next action.
- Open ecosystem. Callers keep their trackers, Git, PRs, CI, cloud agents, merge queues, rollback, and release systems.
Structural constraints
- Core hard dependencies are only
rpi -> {plan, implement, validate}. - Learn and all strategy/factory/specialist skills are off-path or optional.
- Core schemas contain no retry, budget, queue, claim, lease, admission, next-action, closure, release, or delivery state.
- The pure manifest and verdict helpers make no Git, tracker, queue, network, release, or delivery call.
- Deterministic
ao gate checkreports repository check success or failure only. - Semantic validation is a skill responsibility, not a CLI state machine.
Measured learning hypothesis
Collections of durable verdicts may reveal repeated defect classes that deserve better context, tests, or checks. Learn may propose evidence-backed candidates off-path. Promotion requires separate evaluation; no observation silently becomes policy or changes a prior verdict.