Implement

Implement the accepted outcome. Repair ordinary known defects directly. Use the existing intent; no Plan, Recall or Learn worksheet is owed for a clear edit. Implement owns source changes and factual checks. Read changed paths and content identity from Git or ao provenance manifest, not from memory. Copy the author context ID from the session, subagent or thread ID the runtime exposes, or use the identity the caller supplied and label it as supplied. Report it as unknown when neither exists, and never make one up. Copy check results from the command output. ao records these values only as caller-supplied flags and does not derive them.

Rules that decide the result

• Fix the cause, not the oracle. Never loosen an assertion, tolerance, golden, fixture or suppression, or substitute a mock or placeholder, to turn a check green. A check that fails against accepted behavior points at a product defect; changing what the check accepts is an acceptance change and needs caller authority.
• Find live consumers before editing. Search the callers, readers, scripts, docs and tests of every edited function, type, path, format or message. For each one, state whether the change affects it and which check covers it. When retiring or renaming, consumers include installations, lookups and old-name invocations; preserve historical provenance. If the replaced interface never merged or shipped and every consumer update fits the scope, update them and delete the old version in the same change, with no alias or shim; otherwise keep the old version working or report it.
• Name discriminating checks. A behavioral change preserves RED for the expected missing behavior: name the check that fails before the edit and passes after, plus the check for each affected consumer.
• Report only what ran. Give exact commands and results. List every check you did not or could not run as not run. Never call unrun work verified, fixed or green. Output can show as soon as a command runs, so keep secrets out when you build it: credentials stay in the environment, output that may carry tokens goes to a file, and only redacted excerpts are shown.

When changing a public failure path, preserve its native failure signal and check that its recovery advice matches any work already performed. For uncertain write outcomes or unattended callers, use only the relevant checks in Test's error recovery method.

For an accepted bounded shortcut, review feedback or a defect your change introduced, an authorized commit, history rewrite or merge conflict, dependency upgrade, removal, blocked operation, a decision read from another tool's state or your own uncertain write, use only the matching change safety method. These methods preserve existing authority; they grant no new operation.

For authorized service operations, load only the relevant procedure from operations methods (reliability, delivery, incident recovery, resilience, toil); ordinary edits owe no operations phase.

Workflow

1. Read intent, acceptance, scope and repository boundaries before the first write; reuse loaded contracts. RPI boundaries apply when that workflow is explicitly selected. Caller-selected episode tracking follows the session association reference; keep unknowns explicit and invent no parentage or second tracker. Note the revision and files already modified, staged or untracked; report only your own as changed, and ask the caller before editing one holding others' edits.
2. Carry the accepted behavior examples forward unchanged. Use repository domain names in symbols and tests; check observable outcomes through the relevant interface. Run the smallest applicable check before and after editing. Pure refactors, relocations or docs may have an honest green baseline. Prefer existing tests or small discriminating probes.
3. Make the smallest in-scope change. When repairing discovery or checks, preserve the consumer's existing input selection; fixing an error path does not authorize a wider scan. Use a negative control when exclusion matters. Check a representative change against existing constraints before bulk propagation, and the authored source set before broad regeneration. A scripted edit across files first confirms that each anchor matches exactly once; on zero or several matches, write nothing and re-derive it. Repair known failures directly and verify the exact result with a check; a repair does not start another review. A disproved assumption may change the approach within scope; use Plan only for consequential uncertainty.
4. Read the repository's actual check recipe, including instrumentation and environment, rather than reconstructing it from memory. Run targeted tests and lint/static checks before broad integration, and required full checks at integration. Reuse exact-input receipts only while source, tool and relevant environment match. Neither bypass required hooks nor replay a check just to rename its receipt. A required CI job's known failure is actionable before the run ends: repair it within scope, keep the failed subject's evidence and rerun affected checks. Pending jobs do not imply success. A new check, hook or lint rule counts once the aggregate command or CI runs it (a Git hook needs an install path; .git/hooks is not cloned) and it fails on a deliberate violation; else list it under gaps as not wired. Use a local server already on a port only if it provably serves this checkout; stop only processes you started.
5. Refactor while acceptance remains green. Inspect changed tests, fixtures, goldens, tolerances, suppressions and specification text against original intent before handoff. Reread that intent at its source, not a compacted summary, and check the whole delta against it, untracked files included. Remove leftovers of abandoned approaches, temporary probes (one shared marker lets a single search prove none remain) and files your checks made; list any path you cannot explain under gaps as a scope amendment.
6. Have the runtime derive actual changed paths and content identity. A delegated increment awaiting integration returns an exact commit or runtime-derived content digests, author context ID and check facts in the existing handoff. The integrating caller derives subject-manifest.v1 (AgentOps schema schemas/subject-manifest.v1.schema.json) over the complete final subject before judgment; an independently judged increment needs its own manifest. Do not generate both merely because work was delegated.
7. At that boundary, when the repository records AgentOps evidence bindings and changed paths affect bound acceptance evidence, run ao provenance evidence-orphans --root <repo-root> with one --changed <path> per derived path, retain its actual output and refresh affected bindings after repairs. Without ao, list the orphan scan under not run. Never invent or suppress the orphan list.
8. Return the handoff below, then stop.

Handoff

Keep full logs at their source; do not duplicate inventories or status documents.

changed:   <runtime-derived paths>; identity: <commit or content digests>
checks:    <exact command> -> <result>, one per line; before and after for a behavioral change
not run:   <check> -> <reason>; "none" only when every relevant check ran
consumers: <file:line> -> <covering check, or "uncovered">
gaps:      <missing or truncated evidence, known failures, scope amendments needed>
cause:     <defect repairs only: trigger -> mechanism -> symptom at file:line; sites left as follow-up>

Report an uncovered live consumer for a caller scope amendment and continue independent authorized work. Disclose a reversible implementation choice only when its alternative has a material consequence the caller may want to change; name the choice, alternative and reason in the handoff, without a fixed worksheet. Generated companions already included as scope require no new approval; a change to accepted direction, behavior or acceptance always needs caller authority.

Diagnosis, scaffolding and delegated work

For an unexplained failure, first match the reported symptom and reduce the reproduction. Confirm the failing run executes the code you think: revision and local edits, a fresh build rather than a stale artifact or installed binary, the resolved runtime and dependency versions, and required environment values. Ground the leading cause in an observation, not in how the code reads, and name its strongest rival. State one causal prediction, test it with a check the two causes predict differently, and repair the cause supported by the result. Change one thing per probe; when a prediction fails, revert that change, name what ruled the cause out and form a different cause, not a variant of it. Rerun the original scenario. Do not keep collecting hypotheses after the cause is understood. For intermittent verdicts, leaked test state, suspected races, a hung/crashed process, a regression from a known-good state, a failure crossing components or repeated disproved causes, use only the matching experiment in difficult failures. This compact diagnosis path is informed by Matt Pocock's engineering skills.

When scaffolding is the requested change, start from the repository's existing layout and a working vertical slice. See scaffold references only for the relevant tool shape, and generic scaffold examples when the repository has no suitable pattern. Avoid placeholder success paths and a new framework for a one-off operation.

Prefer current-session execution. If delegation is authorized and useful, partition independent writes in isolated workspaces; shared generators and integration serialize. Supply each lane its intent, acceptance, scope and review owner, then integrate its exact content and check facts. A selected wave ends with the caller-requested wave result; do not invent another wave. One fresh review of the integrated candidate can cover unjudged increments; when the integrator owns that review, workers return their handoff without commissioning another. Preserve any separately required lane judgments; a successful process exit is not semantic PASS. Agent Native supplies optional dispatch mechanics.

An explicitly requested one-shot adapter dispatches each supplied operation once, reports its output or error, and stops. Show dispatch count and failure reporting with a dry-run or fixture. It does not silently acquire a scheduler, retry controller or store. Factories require the caller's selection.

Finish

Specialists advise only. Known defects stay implementation work; a genuine causal stall permits at most one bounded fresh helper within caller authority. Respect remaining caller/native bounds and reserve finishing capacity; retries reset neither.

Return facts, not semantic PASS. An implement-only handoff does not authorize Git, tracker or delivery transitions; existing caller authority remains usable. A full outcome request finishes on its checks and CI. It continues to one fresh independent judgment only when the caller asks, a mistake cannot be cheaply undone after it lands, or no deterministic check covers the changed behavior. RPI is optional and explicitly selected. Success is working behavior with usable evidence, not volume of logs or process artifacts.