Generated reader for source-linked technique intelligence.
This reader is a route surface, not technique authority. Authored
TECHNIQUE.md bundles remain stronger than this file, generated JSON,
search output, scout axes, or graph projections.
- Registry version:
1 - Technique entries:
107 - Navigation DAG nodes:
508 - Navigation DAG edges:
507 - Core lens: attention over atomic moves
- Scout axes: family, capability, substrate, execution profile, and risk posture stay non-authoritative
Use Technique Intelligence Guide for the authored contract. Use selection AGENTS and root AGENTS for build, query, and validation command lanes.
The local CLI exposes status, query, explain, and pack actions. Treat their output as a route back to authored bundles, not as execution authority or technique meaning.
| Axis | Values |
|---|---|
| Status | canonical 98, promoted 9 |
| Domain | agent-workflows 57, docs 28, evaluation 12, history 6, system-recovery 3, validation-patterns 1 |
| Kind | artifact 14, assessment 10, composition 7, discovery 2, distribution 4, guardrail 13, handoff 11, ingest 5, lift 12, recovery 6, validation 10, workflow 13 |
| Execution profile hint | medium-agent 21, orchestration-required 53, small-agent 33 |
- technique_intelligence_registry.json
- technique_intelligence_registry.min.json
- technique_intelligence_dag.json
- technique_intelligence_dag.min.json
| Technique | Domain | Kind | Status | Execution hint | Move intent |
|---|---|---|---|---|---|
AOA-T-0001 plan-diff-apply-verify-report |
agent-workflows |
workflow |
canonical |
orchestration-required |
Reduce unsafe, opaque, or non-reviewable agent changes by requiring a visible workflow before and after apply. |
AOA-T-0002 source-of-truth-layout |
docs |
artifact |
canonical |
small-agent |
Reduce documentation drift by giving each recurring kind of project information one canonical document home and one clear update route. |
AOA-T-0003 contract-first-smoke-summary |
evaluation |
validation |
canonical |
small-agent |
Make smoke checks reviewable, automatable, and reusable by treating the machine-readable summary as the default producer-layer contract i... |
AOA-T-0004 intent-plan-dry-run-contract-chain |
agent-workflows |
workflow |
canonical |
medium-agent |
Turn high-level intent into a traceable, reviewable, dry-run-only chain that emits explicit artifacts and a contract verdict before any r... |
AOA-T-0005 new-intent-rollout-checklist |
agent-workflows |
workflow |
promoted |
medium-agent |
Extend an existing intent -> plan -> dry-run -> contract-check pipeline with one new intent_type without losing traceability, artifac... |
AOA-T-0006 latest-alias-plus-history-copy |
evaluation |
artifact |
canonical |
orchestration-required |
Keep machine-readable summaries easy to discover for consumers while also preserving a trustworthy run-by-run history that scanners can a... |
AOA-T-0007 signal-first-gate-promotion |
evaluation |
guardrail |
canonical |
orchestration-required |
Turn an observed validation signal into the default staged path toward a hard gate, so the team can collect evidence, preserve diagnostic... |
AOA-T-0008 published-summary-remediation-snapshot |
evaluation |
lift |
canonical |
orchestration-required |
Create one reviewable remediation snapshot from already-published latest summaries, so operators, dashboards, and agents can see what nee... |
AOA-T-0009 lightweight-status-snapshot |
docs |
artifact |
canonical |
small-agent |
Keep repository entrypoint status easy to scan by turning top-level status docs into short snapshots that summarize the current state and... |
AOA-T-0010 telemetry-integrity-snapshot |
evaluation |
validation |
canonical |
orchestration-required |
Provide one machine-readable diagnostic verdict over published summary invariants, so operators can quickly tell whether telemetry and ar... |
AOA-T-0011 required-vs-optional-source-rendering |
evaluation |
guardrail |
canonical |
orchestration-required |
Keep summary-driven surfaces resilient by enforcing strict behavior for required sources while letting optional enrichments render as `no... |
AOA-T-0012 deterministic-context-composition |
docs |
composition |
canonical |
medium-agent |
Keep agent context scalable by composing it from smaller source fragments into a stable generated artifact whose ordering and source boun... |
AOA-T-0013 single-source-rule-distribution |
docs |
distribution |
canonical |
orchestration-required |
Reduce drift and duplicated maintenance by keeping one canonical rule source and propagating it to multiple agent-facing instruction surf... |
AOA-T-0014 tdd-slice |
agent-workflows |
workflow |
canonical |
orchestration-required |
Reduce ambiguity and regression risk by expressing one bounded behavior slice as tests before implementation, then keeping the implementa... |
AOA-T-0015 contract-test-design |
evaluation |
validation |
canonical |
small-agent |
Reduce breakage at boundaries by expressing the expected interface behavior explicitly and verifying it at the contract surface rather th... |
AOA-T-0016 bounded-context-map |
docs |
artifact |
canonical |
small-agent |
Reduce semantic confusion and misplaced changes by naming bounded contexts, clarifying their responsibilities, and making the handoff int... |
AOA-T-0017 property-invariants |
evaluation |
validation |
canonical |
small-agent |
Reduce weak example-only validation by expressing stable system or domain truths as invariant-oriented checks that should hold across man... |
AOA-T-0018 markdown-technique-section-lift |
docs |
lift |
canonical |
orchestration-required |
Lift stable TECHNIQUE.md section boundaries into a derived section-level surface so retrieval and review can target one bounded section... |
AOA-T-0019 frontmatter-metadata-spine |
docs |
lift |
canonical |
small-agent |
Use shallow frontmatter and derived catalog outputs as a stable metadata spine for bundle identity, routing, and review posture without p... |
AOA-T-0020 evidence-note-provenance-lift |
docs |
lift |
promoted |
small-agent |
Lift typed evidence notes into a bounded provenance surface so readers and tooling can find supporting origin, transfer, or review contex... |
AOA-T-0021 bounded-relation-lift-for-kag |
docs |
lift |
canonical |
small-agent |
Use a small typed set of direct bundle relations as bounded edge hints for generated consumers without pretending that those edges alread... |
AOA-T-0022 risk-and-negative-effect-lift |
docs |
lift |
promoted |
small-agent |
Lift richer Risks language into bounded caution-oriented lookup and reuse so reviewers and later derived surfaces can find failure, har... |
AOA-T-0023 stateless-single-shot-agent |
agent-workflows |
workflow |
canonical |
medium-agent |
Keep quick shell-side agent work bounded to one mostly independent invocation so the fast path stays composable, inspectable, and safe to... |
AOA-T-0024 upstream-mirroring-with-provenance |
docs |
distribution |
canonical |
orchestration-required |
Reuse externally maintained content in a local curated collection without losing origin ownership, attribution, or sync discipline. |
AOA-T-0025 capability-spec-versioning |
docs |
artifact |
canonical |
orchestration-required |
Keep capability behavior reviewable by naming it in one versioned spec instead of letting capability meaning hide inside provider code, r... |
AOA-T-0026 session-capture-as-repo-artifact |
history |
artifact |
canonical |
small-agent |
Preserve AI coding sessions as explicit project-scoped history artifacts so useful reasoning, decisions, and traces stay reviewable after... |
AOA-T-0027 cross-agent-skill-propagation |
docs |
distribution |
canonical |
orchestration-required |
Keep one canonical skill or rule source and propagate it to multiple agent-facing targets so shared guidance stays synchronized without m... |
AOA-T-0028 confirmation-gated-mutating-action |
agent-workflows |
guardrail |
canonical |
small-agent |
Keep the boundary between read or plan work and mutating work explicit so a confirmed action stays reviewable, bounded, and easy to stop... |
AOA-T-0029 nested-rule-loading |
docs |
composition |
canonical |
medium-agent |
Load rule content through a hierarchy of layers so shared defaults, scoped overrides, and explicit precedence can coexist without turning... |
AOA-T-0030 fragmented-agent-context |
docs |
composition |
canonical |
orchestration-required |
Keep agent context authored in smaller bounded fragments so source partitioning, local ownership, and modular review stay clear before an... |
AOA-T-0031 shell-composable-agent-invocation |
agent-workflows |
composition |
canonical |
medium-agent |
Keep agent invocations composable as shell tools so one-shot runs can read from stdin, write to stdout, work with files, and fit into pip... |
AOA-T-0032 context-report-for-ci |
evaluation |
validation |
promoted |
small-agent |
Produce a bounded CI-facing report that tells reviewers how well a composed context covered its expected sources, how much token-estimate... |
AOA-T-0033 decision-rationale-recording |
docs |
artifact |
canonical |
small-agent |
Record one meaningful decision in a compact, reviewable note that preserves context, options, rationale, and consequences without widenin... |
AOA-T-0034 public-safe-artifact-sanitization |
docs |
guardrail |
canonical |
orchestration-required |
Convert potentially sensitive technical material into a shareable artifact that keeps the lesson while removing or generalizing what shou... |
AOA-T-0035 profile-preset-composition |
docs |
composition |
promoted |
medium-agent |
Keep runtime posture reviewable by separating atomic modules, ordered profiles, and named presets into explicit layers instead of hiding... |
AOA-T-0036 render-truth-before-startup |
agent-workflows |
composition |
canonical |
orchestration-required |
Expose the actual composed runtime truth before startup so operators can review the effective service set and rendered config instead of... |
AOA-T-0037 contextual-host-doctor |
evaluation |
validation |
canonical |
orchestration-required |
Expose whether the current host is shaped enough for the selected runtime before startup by running selector-aware readiness checks that... |
AOA-T-0038 one-command-service-lifecycle |
agent-workflows |
workflow |
canonical |
orchestration-required |
Keep a bounded local service stack startable and stoppable through one explicit lifecycle entrypoint so operators do not have to coordina... |
AOA-T-0039 baseline-first-additive-profile-benchmarks |
evaluation |
validation |
canonical |
small-agent |
Keep comparison discipline honest by running one stable baseline profile first, then comparing additive profiles against the same measure... |
AOA-T-0040 skill-vs-command-boundary |
docs |
guardrail |
canonical |
orchestration-required |
Keep reusable capability meaning in a skill artifact and keep user-facing invocation behavior in a command artifact so shared knowledge d... |
AOA-T-0041 skill-marketplace-curation |
docs |
discovery |
canonical |
orchestration-required |
Keep a local skill marketplace or catalog reviewable as a curated discovery layer over upstream-owned skills instead of letting it collap... |
AOA-T-0042 upstream-skill-health-checking |
evaluation |
validation |
promoted |
orchestration-required |
Expose whether an upstream-owned skill source is selectable enough before it reaches a catalog or selector by checking source availabilit... |
AOA-T-0043 multi-source-primary-input-provenance |
docs |
guardrail |
canonical |
small-agent |
Keep the order of source inputs explicit when one bridge or docs surface combines multiple source materials so reviewers can tell which i... |
AOA-T-0044 versionable-session-transcripts |
history |
artifact |
canonical |
orchestration-required |
Turn already-saved AI session history into readable, versionable transcript artifacts so teams can review, compare, cite, hand off, or se... |
AOA-T-0045 witness-trace-as-reviewable-artifact |
history |
artifact |
canonical |
small-agent |
Preserve a nontrivial run as one reviewable witness trace artifact so humans can inspect what the run tried to do, what it actually did,... |
AOA-T-0046 repo-doc-surface-lift |
docs |
lift |
canonical |
orchestration-required |
Lift the repository's bounded public docs/status layer into derived routing knowledge so readers and tooling can answer "which authored r... |
AOA-T-0047 github-review-template-lift |
docs |
lift |
promoted |
orchestration-required |
Lift authored GitHub issue and pull-request review templates into a bounded derived intake surface so reviewers and tooling can find the... |
AOA-T-0048 semantic-review-surface-lift |
docs |
lift |
canonical |
orchestration-required |
Lift authored *_SEMANTIC_REVIEW.md docs into one derived boundary-review surface so readers and tooling can find the current cluster ma... |
AOA-T-0049 dependency-aware-task-graph |
agent-workflows |
workflow |
canonical |
medium-agent |
Represent multi-step coding work as an explicit dependency graph so blocked state, ready state, and ordering decisions stay visible inste... |
AOA-T-0050 ready-work-from-blocker-graph |
agent-workflows |
workflow |
canonical |
medium-agent |
Turn an existing blocker graph into a bounded next-work queue so the next step comes from visible dependency truth rather than from memor... |
AOA-T-0051 commit-triggered-background-review |
agent-workflows |
workflow |
canonical |
medium-agent |
Launch a bounded review pass after a visible commit boundary and preserve the result as a review artifact so findings can be inspected be... |
AOA-T-0052 review-findings-compaction |
agent-workflows |
workflow |
canonical |
medium-agent |
Turn a noisy or repeated findings set into a smaller current review surface by deduping, merging, and revalidating findings against the l... |
AOA-T-0053 local-first-session-index |
history |
artifact |
canonical |
small-agent |
Build a local searchable index over already-saved session artifacts so operators can find and reopen useful history quickly while keeping... |
AOA-T-0054 compaction-resilient-skill-loading |
agent-workflows |
recovery |
canonical |
orchestration-required |
Restore a bounded skill-availability surface after context compaction so the session can rediscover and reload needed skills from canonic... |
AOA-T-0055 requirements-design-tasks-ladder |
agent-workflows |
workflow |
canonical |
medium-agent |
Keep requirements, design, and tasks as visibly separate planning layers so work can move from problem statement to solution shape to exe... |
AOA-T-0056 channelized-agent-mailbox |
agent-workflows |
handoff |
canonical |
small-agent |
Keep agent communication inside durable named channels with ordered replay and explicit acknowledgment so teams can coordinate across ses... |
AOA-T-0057 structured-handoff-before-compaction |
agent-workflows |
handoff |
canonical |
small-agent |
Write one structured handoff artifact before compaction or session rollover so the next session can resume from an explicit continuation... |
AOA-T-0058 receipt-confirmed-handoff-packet |
agent-workflows |
handoff |
promoted |
small-agent |
Require the receiving side to record explicit receipt of a handoff packet before continuing work so transfer of responsibility is visible... |
AOA-T-0059 git-verified-handoff-claims |
agent-workflows |
handoff |
promoted |
small-agent |
Verify concrete handoff claims against visible git state before continuation so the next session trusts reviewable repo evidence rather t... |
AOA-T-0060 session-opening-ritual-before-work |
agent-workflows |
handoff |
canonical |
orchestration-required |
Start a resumed or handed-off session with one visible read-and-verify ritual before the first mutation so work begins from current state... |
AOA-T-0061 cross-repo-resource-map-bootstrap |
agent-workflows |
handoff |
canonical |
small-agent |
Bootstrap cross-repo work from one explicit resource map so the next session can see which repositories and surfaces matter before deeper... |
AOA-T-0062 episode-bounded-agent-loop |
agent-workflows |
handoff |
canonical |
orchestration-required |
Break longer work into explicit episodes with checkpoints and continue, stop, or escalate decisions so continuation stays reviewable inst... |
AOA-T-0063 versioned-agent-registry-contract |
docs |
artifact |
canonical |
orchestration-required |
Keep registry-facing capability publication reviewable by expressing entries as named versioned records with explicit references and meta... |
AOA-T-0064 capability-discovery |
docs |
discovery |
canonical |
orchestration-required |
Keep capability discovery reviewable by expressing lookup as one explicit bounded query contract over already-published registry entries... |
AOA-T-0065 mcp-gateway-proxy |
agent-workflows |
composition |
canonical |
orchestration-required |
Front multiple configured MCP servers through one bounded gateway proxy so callers can use one reviewable tool surface with explicit meta... |
AOA-T-0066 transcript-replay-artifact |
history |
artifact |
canonical |
small-agent |
Turn already-saved session history into a replayable artifact so reviewers can inspect message flow and timeline without reopening captur... |
AOA-T-0067 transcript-linked-code-lineage |
history |
artifact |
canonical |
small-agent |
Link code history back to saved session evidence so reviewers can reopen the originating transcript or rationale without widening the bun... |
AOA-T-0068 fail-closed-evidence-gate |
agent-workflows |
guardrail |
canonical |
orchestration-required |
Stop mutating execution at the boundary unless an explicit allow verdict exists, and emit reviewable evidence for blocked or allowed path... |
AOA-T-0069 approval-bound-durable-jobs |
agent-workflows |
handoff |
canonical |
small-agent |
Keep longer-running work durable across one explicit approval seam so checkpoint, pause, and resume remain reviewable without widening in... |
AOA-T-0070 two-stage-document-ocr-pipeline |
agent-workflows |
ingest |
canonical |
orchestration-required |
Turn screenshots, scans, or document images into a bounded OCR handoff by keeping text-region detection or layout analysis separate from... |
AOA-T-0071 template-backed-field-extraction-after-ocr |
agent-workflows |
ingest |
canonical |
orchestration-required |
Turn OCR handoff text into a bounded field object such as merchant, date, amount, and currency through explicit templates, heuristics, an... |
AOA-T-0072 perceptual-media-dedupe-with-threshold-review |
agent-workflows |
ingest |
canonical |
orchestration-required |
Make screenshot, meme, or photo deduplication reviewable by grouping near-duplicate media through perceptual similarity while keeping unc... |
AOA-T-0073 semantic-media-bucketing-with-vision-plus-ocr |
agent-workflows |
ingest |
canonical |
orchestration-required |
Provide one bounded classification route for mixed media by combining visual semantics with OCR side text so screenshots, memes, receipts... |
AOA-T-0074 telegram-export-normalization-to-local-store |
agent-workflows |
ingest |
canonical |
orchestration-required |
Turn Telegram export or sync output into a bounded, provenance-preserving local store so later workflows can inspect, route, or distill m... |
AOA-T-0075 session-donor-harvest |
agent-workflows |
lift |
canonical |
small-agent |
Turn one reviewed session artifact into a bounded donor pack of reusable candidates so later placement, drafting, and promotion work star... |
AOA-T-0076 owner-layer-triage |
agent-workflows |
assessment |
canonical |
medium-agent |
Take one bounded reusable unit and assign it to one primary owner layer, one smallest honest next artifact, and one rejected nearest-wron... |
AOA-T-0077 harvest-packet-contract |
agent-workflows |
handoff |
canonical |
small-agent |
Keep one bounded HARVEST_PACKET contract over a reviewed session artifact so later family seams can consume one explicit packet shape i... |
AOA-T-0078 decision-fork-cards |
agent-workflows |
assessment |
canonical |
orchestration-required |
Turn one reviewed session or harvest packet into explicit decision fork cards so materially different next routes stay legible and review... |
AOA-T-0079 risk-passport-lift |
agent-workflows |
assessment |
canonical |
orchestration-required |
Attach one small risk passport to each explicit next route so route choice keeps bounded control posture, difficulty, and stop-condition... |
AOA-T-0080 session-drift-taxonomy |
agent-workflows |
assessment |
canonical |
medium-agent |
Classify repeated post-session friction into bounded drift types so a later diagnosis pass can describe what kind of problem is present b... |
AOA-T-0081 diagnosis-from-reviewed-evidence |
agent-workflows |
assessment |
canonical |
medium-agent |
Turn reviewed friction evidence into a bounded diagnosis packet so symptoms, probable causes, owner hints, and explicit unknowns become l... |
AOA-T-0082 repair-shape-from-diagnosis |
agent-workflows |
recovery |
canonical |
orchestration-required |
Turn a reviewed diagnosis packet into the smallest honest repair shape so the next step stays bounded and explicit instead of becoming va... |
AOA-T-0083 checkpoint-bound-self-repair |
agent-workflows |
recovery |
canonical |
orchestration-required |
Keep self-repair behind explicit checkpoint posture so bounded repair stays reviewable and reversible instead of feeling like silent self... |
AOA-T-0084 progression-evidence-lift |
agent-workflows |
lift |
canonical |
small-agent |
Lift reviewed session evidence into a bounded multi-axis progression delta so growth stays legible and evidence-backed without inventing... |
AOA-T-0085 multi-axis-quest-overlay |
agent-workflows |
lift |
canonical |
orchestration-required |
Add quest-, RPG-, or chronicle-shaped reflection to a bounded progression or route-reading surface so the path becomes easier to read wit... |
AOA-T-0086 automation-fit-matrix |
agent-workflows |
assessment |
canonical |
medium-agent |
Classify one recurring manual route with a small automation-fit matrix so the result says whether the route is honestly ready for automat... |
AOA-T-0087 human-loop-to-first-landing |
agent-workflows |
assessment |
canonical |
medium-agent |
Lift one recurring human loop into the first honest automation-facing landing without skipping the difference between a bounded skill, a... |
AOA-T-0088 approval-sensitivity-check |
agent-workflows |
assessment |
canonical |
orchestration-required |
Classify whether one automation candidate crosses approval, rollback, hidden authority, or self-change boundaries so later seed or repair... |
AOA-T-0089 quest-unit-promotion-review |
agent-workflows |
assessment |
canonical |
medium-agent |
Review one repeated, reviewed quest-shaped unit and decide whether it should stay a quest or be promoted into the next honest owner surface. |
AOA-T-0090 nearest-wrong-target-rejection |
agent-workflows |
guardrail |
canonical |
small-agent |
Make the nearest wrong target explicit when writing a bounded owner or promotion verdict so adjacent layers stay visibly distinct under r... |
AOA-T-0091 workspace-root-ingress-and-mutation-gate |
agent-workflows |
guardrail |
canonical |
small-agent |
Enter federated workspaces through one explicit ingress pass and require one explicit guard pass before risky mutation so session posture... |
AOA-T-0092 audit-to-closeout-proof-loop |
agent-workflows |
workflow |
canonical |
medium-agent |
Turn a reviewed audit finding set into a proof-backed closeout loop by requiring live confirmation of each finding, a bounded owner-surfa... |
AOA-T-0093 recommendation-truth-vs-host-actionability |
agent-workflows |
guardrail |
canonical |
orchestration-required |
Keep router recommendation truth and host actionability separate so control-plane reports stay honest, activation only reflects what the... |
AOA-T-0094 canonical-owner-with-validated-mirror |
docs |
distribution |
canonical |
orchestration-required |
Keep one canonical owner for a cross-repo contract while still permitting nearby mirrors, as long as each mirror proves exact parity inst... |
AOA-T-0095 github-only-owner-endcap-with-reality-sync |
agent-workflows |
workflow |
canonical |
medium-agent |
Close a bounded remote-only owner surface through GitHub-native issue and PR flow, then immediately sync the coordination or seed-garden... |
AOA-T-0096 pinned-validation-matrix-before-generated-publish |
agent-workflows |
validation |
canonical |
orchestration-required |
Rebuild and validate generated outputs against the same workflow-pinned sibling refs, mirrored surfaces, or bounded upstream inputs that... |
AOA-T-0097 degrade-reground-recover |
system-recovery |
recovery |
canonical |
orchestration-required |
Keep partial failure honest and useful by moving through one bounded sequence: degrade safely, reground on stronger authority, and recove... |
AOA-T-0098 receipt-first-failure-analysis |
validation-patterns |
validation |
canonical |
orchestration-required |
Turn failure review into bounded evidence-led iteration by starting from source-owned receipts, separating facts from guesses, and defini... |
AOA-T-0099 isolated-service-stop-on-shared-substrate |
system-recovery |
recovery |
canonical |
orchestration-required |
Stop one bounded service in a shared local stack without needlessly tearing down the substrate that other verification, recovery, or foll... |
AOA-T-0100 stress-receipt-reground-closeout |
system-recovery |
recovery |
canonical |
orchestration-required |
Provide a reusable workflow for bounded chaos or stress events where the right move is not "repair immediately", but: - record the stress... |
AOA-T-0101 local-pattern-adoption-gate |
agent-workflows |
guardrail |
canonical |
orchestration-required |
Gate one shared practice pattern before it becomes durable local behavior, so approval, precedent, or repeated usefulness does not silent... |
AOA-T-0102 skill-proposal-handoff-packet |
agent-workflows |
handoff |
canonical |
small-agent |
Turn one technique-side adoption pressure into a bounded skill-proposal handoff packet, so the receiving skill owner gets enough trigger,... |
AOA-T-0103 adopted-practice-retention-review |
agent-workflows |
assessment |
canonical |
medium-agent |
Review one already-adopted practice to decide whether it should remain active, be revised, be quarantined, be deferred for another review... |
AOA-T-0104 superseded-practice-obsolescence-route |
agent-workflows |
handoff |
canonical |
orchestration-required |
Route one adopted or shadowed practice that should not remain active as-is toward supersession, merge, reanchor, defer, drop, or deprecat... |
AOA-T-0105 single-missing-evidence-request |
agent-workflows |
guardrail |
canonical |
orchestration-required |
Ask for exactly one missing evidence object that could change a bounded review state, so review can narrow without broad research, verdic... |
AOA-T-0106 single-scoped-evidence-reference |
docs |
artifact |
canonical |
small-agent |
Offer exactly one scoped evidence reference with relevance, support limit, and reliance condition so review can use a source without trea... |
AOA-T-0107 single-locus-claim-challenge |
agent-workflows |
guardrail |
canonical |
orchestration-required |
Challenge exactly one claim at one vulnerable locus, naming pressure reason and next support question so review can apply pressure withou... |
Use the registry to choose, compare, explain, or pack one move. Leave this repo when the object becomes execution workflow, proof verdict, dispatch policy, runtime behavior, role contract, scenario composition, memory writeback, or graph semantics.