AI Agent Workflow Collection

Route AoA decision-lane work through the workspace decision graph first, checking status/issues and the smallest impact packet before choosing find, create, or correct while keeping repo-local decision files authoritative. Use when the task mentions docs/decisions, decision indexes, decision graph, ADRs, changed paths, source-surface rationale, supersession, or cross-repo decision symmetry. Do not use for ordinary docs edits, unresolved source-of-truth mapping, or trivial changes that do not need durable rationale.

Correct, supersede, reindex, or repair graph-coverage drift in AoA decision records by checking graph issues, editing source notes first, rebuilding repo-local indexes, and refreshing the workspace decision graph. Use when a decision note, index metadata, source-surface list, status, supersession, generated index, unknown surface, or graph packet is stale or wrong. Do not use for pure lookup or for creating a new decision note.

Create a new AoA decision record in the owning repository by checking graph status/issues, using impact packets for placement, and following repo-local decision law for the actual source note and indexes. Use when a real structural, workflow, tooling, source/export, MCP, skill, or authority decision needs durable rationale. Do not use when no decision has been made, the change is trivial, the target repo has unresolved graph issues, or an existing note only needs correction.

Find AoA decision records through workspace graph search and impact packets, then verify claims against repo-local source notes and generated decision indexes. Use when prior decision rationale, changed paths, owner surfaces, source surfaces, status, or supersession links are needed before work continues. Do not use when the task is to create or edit a decision record.

Route AoA eval-lane work by raising the available session readiness packet, checking existing local and central eval surfaces first, including any available repo-family or workspace eval-port inventory with missing/skeleton/active/invalid status, then choosing apply, local intake need, local design, or session mining while preserving owner boundaries. Use when a task asks whether an eval exists, whether one should be added, how local evals connect to aoa-evals, how local eval-port status should route work, or how validators, tests, scripts, traces, regressions, trigger misses, local eval ports, aoa-evals-mcp, candidate packets, or per-session eval readiness should be used as evaluation surfaces. Do not use for ordinary one-off tests, source-of-truth mapping, decision records, memo writeback, or direct central aoa-evals proof-authority edits.

Run or route an already selected eval, validator, test, or script, then report command results, artifacts, generated drift, proof limits, and next route. Use when the eval surface is already selected and the task is to apply it rather than choose or design it. Do not use when no existing eval has been selected, when an intake need should be recorded first, or when central proof promotion is being requested.

Design a bounded local eval suite, prompt set, report, or validator plan after selection finds no adequate existing surface, using deterministic checks first and trace or rubric review only when needed. Use when a local eval needs positive, negative, collision, or regression cases around a clear behavior or invariant. Do not use when an existing eval should simply run, when only a local intake packet is needed, or when central aoa-evals proof doctrine is the target.

Create a repo-local eval need packet under evals/intake when no existing eval fits, preserving owner route, evidence refs, rejected alternatives, proposed checks, and proof limits. Use when a repeated failure or missing behavior needs local eval pressure without central proof acceptance. Do not use when an existing eval can be applied, when the missing object is a memo or decision, or when the user asks for central aoa-evals acceptance.

Inspect repo-local eval ports, tests, validators, scripts, reports, intake files, and central aoa-evals standards before new eval work, then select the smallest existing surface or a no-fit route. Use when the task asks which eval should apply to a change, failure, regression, or trigger miss. Do not use when the eval is already selected and should run, when no-fit has already been established, or when the real issue is source authority rather than eval coverage.

Mine .aoa session evidence for missed eval triggers, candidate regression cases, and local eval pressure only after web research and repo owner surfaces are checked, returning raw and segment refs with freshness rather than proof. Use when the user asks to study sessions for eval moments, missed triggers, repeated failures, or session-derived eval cases. Do not use for raw .aoa preservation, hook repair, search-index maintenance, memo writeback, decision records, or treating session history as reviewed proof.

Bridge session evidence into memory writeback decisions without turning raw session history into reviewed memory. Use when a live session, closeout, PR, commit, or review reveals a memory-worthy decision, failure, route change, owner-boundary event, service contract, eval posture, or consumer handoff and the agent needs to decide whether to write a local memo candidate, prepare a reviewed-intake export, or record an explicit no-writeback stop line. Do not use for raw session capture, reviewed memory landing itself, proof verdicts, eval scoring, or generic summaries with no owner route.

Route OS Abyss ABI, provenance, signatures, SBOM, SLSA/in-toto, Sigstore/Cosign, C2PA, TUF, SCITT, durable evidence, drift, producer profiles, and trust-gate work through existing abyss-machine read models and owner-local producer routes. Use when an agent needs to inspect or consume bundles, containers, models, runtimes, reports, media exports, source seeds, portable memory bundles, browser extensions, generated machine surfaces, or release artifacts without creating a second trust MCP or violating owner boundaries. Do not use for ordinary local edits, raw .aoa session search, or direct signing/promotion/registry mutation without an authorized owner route.