| 1 |
Name a Skill and describe when to use it |
Skill authoring |
Located source instance |
| 2 |
Phased/Stepped Execution |
Workflow control |
Located source instance |
| 3 |
Workflow Mode Branching |
Workflow control |
Located source instance |
| 4 |
$ARGUMENTS Variable |
Skill authoring |
Source unconfirmed |
| 5 |
Persona/Role Assignment |
Prompt design |
Source unconfirmed |
| 6 |
Negative Constraints |
Prompt design |
Located source instance |
| 7 |
Interactive Flow Control |
Workflow control |
Located source instance |
| 8 |
Confirmation Gates |
Safety & trust |
Located source instance |
| 9 |
Progress Feedback |
Workflow control |
Located source instance |
| 10 |
Prompt Injection Defense |
Safety & trust |
Located source instance |
| 11 |
Sensitive Data Redaction |
Safety & trust |
Located source instance |
| 12 |
Read-Only Boundary |
Safety & trust |
Located source instance |
| 13 |
Activation Scope |
Skill authoring |
Located source instance |
| 14 |
Structured Output Templates |
Output contracts |
Located source instance |
| 15 |
Error Handling / Degradation |
Workflow control |
Source unconfirmed |
| 16 |
Configuration Persistence |
Context management |
Source unconfirmed |
| 17 |
Cross-Platform Handling |
Tool use |
Source unconfirmed |
| 18 |
Multi-Agent Orchestration |
Agent orchestration |
Located source instance |
| 19 |
Skill Composition |
Skill authoring |
Source unconfirmed |
| 20 |
Intent Classification |
Workflow control |
Located source instance |
| 21 |
Tool Routing Tables |
Tool use |
Source unconfirmed |
| 22 |
Deduplication / Consensus |
Agent orchestration |
Source unconfirmed |
| 23 |
Reference File Injection |
Context management |
Located source instance |
| 24 |
Domain Knowledge Embedding |
Context management |
Source unconfirmed |
| 25 |
Few-Shot Examples |
Prompt design |
Located source instance |
| 26 |
Evidence Chain / Proof-of-Work |
Evaluation & feedback |
Located source instance |
| 27 |
Scoring Rubrics |
Evaluation & feedback |
Located source instance |
| 28 |
Self-Critique |
Evaluation & feedback |
Located source instance |
| 29 |
Feedback Solicitation |
Workflow control |
Source unconfirmed |
| 30 |
Version Check |
Tool use |
Source unconfirmed |
| 32 |
Hub-and-Spoke SDLC State Machine |
Agent orchestration |
Located source instance |
| 34 |
Dual-Model Adversarial Planning |
Evaluation & feedback |
Located source instance |
| 35 |
Cost-Optimized Model Routing |
Agent orchestration |
Located source instance |
| 36 |
Handoff Context Protocol |
Context management |
Located source instance |
| 41 |
Loop Prevention with Max Iterations |
Workflow control |
Located source instance |
| 42 |
Agent Memory Isolation |
Agent orchestration |
Located source instance |
| 43 |
Isolated Review Checkout |
Tool use |
Located source instance |
| 47 |
Evidence-First Review |
Evaluation & feedback |
Located source instance |
| 48 |
Rule-Catalog Review (YAML) |
Evaluation & feedback |
Source unconfirmed |
| 50 |
Adversarial Triad + Counterarguments |
Agent orchestration |
Located source instance |
| 51 |
Schema Validation Gate |
Output contracts |
Located source instance |
| 52 |
Rubric-Based Model Evaluation |
Evaluation & feedback |
Located source instance |
| 59 |
Rescue-Tag-Before-Destructive-Operation |
Safety & trust |
Source unconfirmed |
| 60 |
Tiered Permission Model (RED/DEFER/GREEN) |
Safety & trust |
Source unconfirmed |
| 61 |
Data Classification Matrix (4-Level) |
Safety & trust |
Source unconfirmed |
| 67 |
Whole-Payload Skill Review |
Safety & trust |
Located source instance |
| 68 |
Orchestrator-Only (No Direct Data) |
Safety & trust |
Source unconfirmed |
| 70 |
State File as Sole Continuity |
Context management |
Located source instance |
| 72 |
Pull-Based Kanban Orchestration |
Agent orchestration |
Source unconfirmed |
| 73 |
Resumable Idempotent Actions |
Workflow control |
Located source instance |
| 74 |
Autonomous PR Feedback Resolution |
Workflow control |
Located source instance |
| 76 |
Staggered Burst Query + Rate Limits |
Tool use |
Source unconfirmed |
| 77 |
Time-Boxed Investigation |
Workflow control |
Source unconfirmed |
| 81 |
NL to Relational Schema Decomposition |
Prompt design |
Source unconfirmed |
| 83 |
Audience-Purpose Content Calibration |
Output contracts |
Located source instance |
| 87 |
Eager Incremental Materialization |
Workflow control |
Located source instance |
| 89 |
Writability Rules + Substitution Tables |
Output contracts |
Source unconfirmed |
| 92 |
DAG Journey with Typed Gates |
Workflow control |
Source unconfirmed |
| 94 |
Promise Detection and KB Sync |
Output contracts |
Source unconfirmed |
| 95 |
Mandatory Self-Learning After Failure |
Context management |
Located source instance |
| 96 |
Risk-Ordered Batch Migration + Build-Verify |
Workflow control |
Source unconfirmed |
| 100 |
Progressive Disclosure Architecture |
Context management |
Located source instance |
| 101 |
Creative Philosophy Scaffolding |
Prompt design |
Located source instance |
| 103 |
Reconnaissance-Then-Action |
Tool use |
Located source instance |
| 104 |
Helper Script as Black Box |
Tool use |
Located source instance |
| 106 |
Section-by-Section Collaborative Drafting |
Workflow control |
Located source instance |
| 107 |
Reader Testing with Sub-Agent |
Evaluation & feedback |
Located source instance |
| 108 |
Blind A/B Comparison |
Evaluation & feedback |
Located source instance |
| 109 |
Composio 3-Step SaaS Integration |
Tool use |
Located source instance |
| 115 |
QA Sub-Agent with Visual Verification |
Evaluation & feedback |
Located source instance |
| 116 |
Preserve Computation in Editable Artifacts |
Output contracts |
Located source instance |
| 119 |
Eval-Driven Skill Improvement Loop |
Skill authoring |
Located source instance |
| 121 |
Typed Memory Taxonomy |
Context management |
Source unconfirmed |
| 123 |
Scoped Authorization and Action Risk |
Safety & trust |
Located source instance |
| 127 |
Parallel-Safe Step Identification |
Tool use |
Located source instance |
| 128 |
Context Compaction Survival |
Context management |
Located source instance |
| 132 |
Hook-Driven Automation |
Tool use |
Source unconfirmed |
| 133 |
Compositional Prompt Assembly |
Skill authoring |
Located source instance |
| 134 |
Tool-Constraint Boundaries |
Safety & trust |
Located source instance |
| 135 |
Fork vs Fresh Spawning |
Agent orchestration |
Source unconfirmed |
| 141 |
REPL as Tool Composition |
Tool use |
Source unconfirmed |
| 143 |
Pre-Set Catalog with Generate-On-The-Fly Fallback |
Workflow control |
Located source instance |
| 144 |
Agent-Centric Tool Design Principles |
Tool use |
Located source instance |
| 145 |
Evidence-Bound Verification Rule |
Prompt design |
Located source instance |
| 146 |
Rationalization-Prevention Table |
Prompt design |
Located source instance |
| 148 |
Anti-Performative-Agreement Vocabulary Ban |
Workflow control |
Located source instance |
| 149 |
Pressure-Scenario TDD for Skill Authoring |
Skill authoring |
Located source instance |
| 150 |
Continuous Execution Mandate |
Workflow control |
Located source instance |
| 151 |
Explicit Preconditions with Gate Markers |
Prompt design |
Located source instance |
| 152 |
DOT-Graph Decision Flow Embedded in Prompt |
Workflow control |
Located source instance |
| 153 |
Marketplace Source-Type Polymorphism (SHA-Pinned) |
Tool use |
Source unconfirmed |
| 154 |
Self-Looping Stop Hook (Ralph Loop) |
Workflow control |
Source unconfirmed |
| 155 |
Capability Retirement with Migration Paths |
Skill authoring |
Located source instance |
| 156 |
No-Pre-Judging Reviewer Dispatch |
Agent orchestration |
Located source instance |
| 158 |
Plan-Mandated Defect Is Still a Finding |
Evaluation & feedback |
Located source instance |
| 159 |
Diff-Scoped Review with Named-Risk Escape Hatch |
Evaluation & feedback |
Located source instance |
| 160 |
Reviewer Does Not Re-Run the Implementer’s Tests |
Evaluation & feedback |
Located source instance |
| 164 |
Batch Remediation: One Fixer for All Findings |
Agent orchestration |
Located source instance |
| 165 |
Reviewer-Gate Task Right-Sizing |
Workflow control |
Located source instance |
| 166 |
Verbatim Global-Constraints Block |
Context management |
Located source instance |
| 167 |
Per-Task Interfaces Block (Consumes/Produces) |
Agent orchestration |
Located source instance |
| 170 |
Just-In-Time Capability Offer |
Workflow control |
Source unconfirmed |
| 171 |
Action-Verb Tool Abstraction |
Skill authoring |
Located source instance |
| 173 |
Capability-Absent Anti-Fabrication Fallback |
Tool use |
Located source instance |
| 174 |
Reachability-Scoped Delta |
Evaluation & feedback |
Source unconfirmed |
| 175 |
Tool-Ran Tri-State Provenance |
Evaluation & feedback |
Located source instance |
| 176 |
Differential Baseline-Equivalence Oracle |
Evaluation & feedback |
Located source instance |
| 177 |
Capability Smoke-Test Over Presence-Check |
Tool use |
Located source instance |
| 178 |
Derivation-Order Staleness Check |
Context management |
Located source instance |
| 181 |
Cross-Service Credential-Routing Detection |
Safety & trust |
Source unconfirmed |
| 183 |
Untrusted-Content-as-Markup Escaping |
Safety & trust |
Located source instance |
| 188 |
Immutable-Slug Identity with Rename Redirect Map |
Tool use |
Source unconfirmed |
| 190 |
Invocation Economics (Context vs Cognitive Load) |
Skill authoring |
Located source instance |
| 192 |
No-Op Test (Behavior-vs-Default Pruning) |
Evaluation & feedback |
Located source instance |
| 193 |
Two-Axis Completion Criterion (Checkable + Exhaustive) |
Workflow control |
Located source instance |
| 194 |
Controlled Vocabulary with Banned Synonyms |
Context management |
Located source instance |
| 197 |
Fog of War (Frontier Planning) |
Workflow control |
Located source instance |
| 198 |
Two-Axis Orthogonal Review (No Cross-Axis Reranking) |
Evaluation & feedback |
Located source instance |
| 201 |
Absence-of-Seam as a Finding |
Evaluation & feedback |
Located source instance |
| 202 |
Feedback-Loop-Before-Hypothesis Gate |
Workflow control |
Located source instance |
| 204 |
Static-Trace Verification When Execution Is Unsafe |
Evaluation & feedback |
Located source instance |
| 206 |
ADR-Worthiness Triple Gate |
Workflow control |
Located source instance |
| 207 |
Wait for a job’s terminal state |
Workflow control |
Located source instance |
| 208 |
Validate measured artifact constraints |
Evaluation & feedback |
Located source instance |
| 209 |
Checkpoint partial bulk progress |
Workflow control |
Located source instance |
| 210 |
Use delta tokens for recurring synchronization |
Context management |
Located source instance |
| 211 |
Recover according to failure class |
Tool use |
Located source instance |
| 212 |
Read fresh version tokens before update |
Tool use |
Located source instance |
| 213 |
Manage helper process lifetime |
Workflow control |
Located source instance |
| 214 |
Choose capabilities from current metadata |
Tool use |
Located source instance |
| 215 |
Prove paginated retrieval is complete |
Tool use |
Located source instance |
| 216 |
Validate per-item outcomes inside batch success |
Tool use |
Located source instance |
| 217 |
Materialize expiring outputs before handoff |
Workflow control |
Located source instance |
| 218 |
Pilot a fixed extraction contract |
Evaluation & feedback |
Located source instance |
| 219 |
Check replace semantics before editing |
Tool use |
Located source instance |
| 220 |
Resolve opaque IDs within their scope |
Tool use |
Located source instance |
| 221 |
Keep an undo ledger for file operations |
Safety & trust |
Located source instance |
| 222 |
Resolve schema references before arguments |
Tool use |
Located source instance |
| 223 |
Bind actions to the selected tenant |
Safety & trust |
Located source instance |
| 224 |
Explicit continuation contract |
Context management |
Located source instance |
| 225 |
Identify the actor before inferring intent |
Context management |
Located source instance |
| 226 |
Abuse cases beside intended use cases |
Evaluation & feedback |
Located source instance |
| 227 |
Grade the declared artifact kind |
Evaluation & feedback |
Located source instance |
| 228 |
Validate kill assumptions before secondary preferences |
Workflow control |
Located source instance |
| 229 |
Atomic claim with payload binding |
Agent orchestration |
Located source instance |
| 230 |
Cache keys encode response-changing inputs |
Safety & trust |
Located source instance |
| 231 |
Resource freshness versus reading relevance |
Context management |
Located source instance |
| 232 |
Place checks by feedback cost |
Workflow control |
Located source instance |
| 233 |
Validate the consumer interpretation |
Evaluation & feedback |
Located source instance |
| 234 |
Resolve confinement and ownership before destruction |
Safety & trust |
Located source instance |
| 235 |
Exemption authority outside the constrained unit |
Safety & trust |
Located source instance |
| 236 |
Retain rejected experiment evidence |
Context management |
Located source instance |
| 237 |
Bind grader results to the full expectation set |
Evaluation & feedback |
Located source instance |
| 238 |
Review changes to the quality bar itself |
Evaluation & feedback |
Located source instance |
| 239 |
Choose interaction by investigation shape |
Agent orchestration |
Located source instance |
| 240 |
Reference closure in the received artifact |
Context management |
Located source instance |
| 241 |
Separate workflow, perspective and entry point |
Agent orchestration |
Located source instance |
| 242 |
Measured baseline ratchet |
Evaluation & feedback |
Located source instance |
| 243 |
Preserve measurement source kinds |
Output contracts |
Located source instance |
| 244 |
Noise-aware intervention comparison |
Evaluation & feedback |
Located source instance |
| 245 |
Owned negative routing controls |
Evaluation & feedback |
Located source instance |
| 246 |
One active routing authority |
Context management |
Located source instance |
| 247 |
Protected content through an editable projection |
Context management |
Located source instance |
| 248 |
Question-to-signal observability |
Tool use |
Located source instance |
| 249 |
Ordered reconciliation of reviewer findings |
Evaluation & feedback |
Located source instance |
| 250 |
Pass review artifacts through inert input channels |
Safety & trust |
Located source instance |
| 251 |
Task acceptance and standing readiness |
Evaluation & feedback |
Located source instance |
| 252 |
Reuse verification until its assumptions change |
Evaluation & feedback |
Located source instance |
| 253 |
Verify the evidence channel itself |
Evaluation & feedback |
Located source instance |
| 254 |
Use a deadline independent of stream activity |
Tool use |
Located source instance |
| 255 |
Keep answer keys outside agent reach |
Evaluation & feedback |
Located source instance |
| 256 |
Keep attempts, grades and traces coherent |
Evaluation & feedback |
Located source instance |
| 257 |
Separate build variance from scoring variance |
Evaluation & feedback |
Located source instance |
| 258 |
Preserve the complete protocol envelope |
Context management |
Located source instance |
| 259 |
Concrete, intent-aware verification prompts |
Evaluation & feedback |
Located source instance |
| 260 |
Retain full validation beside a compatible schema |
Tool use |
Located source instance |
| 261 |
Preserve coordinate units through tool handoffs |
Tool use |
Located source instance |
| 262 |
Regrade all variants after a criterion change |
Evaluation & feedback |
Located source instance |
| 263 |
Separate definition versions from runs |
Workflow control |
Located source instance |
| 264 |
Move deterministic loop steps into code |
Workflow control |
Located source instance |
| 265 |
Use fresh no-change controls after drift |
Evaluation & feedback |
Located source instance |
| 266 |
Grade the environment left by an agent |
Evaluation & feedback |
Located source instance |
| 267 |
Freeze comparison references across rounds |
Evaluation & feedback |
Located source instance |
| 268 |
Reconnect event streams without gaps |
Tool use |
Located source instance |
| 269 |
Record the origin of evaluation references |
Evaluation & feedback |
Located source instance |
| 270 |
Keep selection feedback out of revision prompts |
Evaluation & feedback |
Located source instance |
| 271 |
Classify text by its control function |
Skill authoring |
Located source instance |
| 272 |
Intent-matched, fresh catalog recommendations |
Context management |
Located source instance |
| 273 |
Confirm the selected configuration as a whole |
Evaluation & feedback |
Located source instance |
| 274 |
Classify the failing layer when optimization stalls |
Evaluation & feedback |
Located source instance |
| 275 |
Evaluate the real application entry point |
Evaluation & feedback |
Located source instance |
| 276 |
Keep pruning assumptions and deferred choices |
Workflow control |
Located source instance |
| 277 |
Check instruction scope before resolving conflicts |
Context management |
Located source instance |
| 278 |
Assemble cached prefixes by stability |
Context management |
Located source instance |
| 279 |
Validate the whole streamed turn before acting |
Tool use |
Located source instance |
| 280 |
Promote opaque actions to typed control points |
Tool use |
Located source instance |
| 281 |
Trace invalid data back to its origin |
Evaluation & feedback |
Located source instance |
| 282 |
Prove behavioral RED before minimal GREEN |
Workflow control |
Located source instance |
| 283 |
Establish counter semantics before totals |
Evaluation & feedback |
Located source instance |
| 284 |
Calibrate instruction specificity to fragility |
Skill authoring |
Located source instance |
| 285 |
Wait on a fresh observable predicate |
Tool use |
Located source instance |
| 286 |
Use an independent, mutation-sensitive oracle |
Evaluation & feedback |
Located source instance |
| 287 |
Disclose consequential assumptions before acting |
Workflow control |
Located source instance |
| 288 |
Implement only presently required complexity |
Prompt design |
Located source instance |
| 289 |
Reformulate tasks as observable goals |
Workflow control |
Located source instance |
| 290 |
Make request-traceable surgical edits |
Safety & trust |
Located source instance |
| 291 |
Label implementation approximation provenance |
Output contracts |
Located source instance |
| 292 |
Lock invariants, vary composition |
Context management |
Located source instance |
| 293 |
Contextual rule precedence |
Prompt design |
Located source instance |
| 294 |
Successor screens require an action and carried state |
Evaluation & feedback |
Located source instance |
| 295 |
Provenance for precise-looking mock content |
Safety & trust |
Located source instance |
| 296 |
Behavioral motivation before adding effects |
Prompt design |
Located source instance |
| 297 |
Match reference granularity to inspection needs |
Context management |
Located source instance |
| 298 |
Audit and preserve redesign contracts |
Workflow control |
Located source instance |
| 299 |
Semantic token handoff contract |
Context management |
Located source instance |
| 300 |
Inspect the specimen before translating it |
Workflow control |
Located source instance |
| 301 |
Remove presentation answer cues |
Evaluation & feedback |
Located source instance |
| 302 |
Guardrail negative-control proof |
Evaluation & feedback |
Located source instance |
| 303 |
Deletion probe before abstraction |
Prompt design |
Located source instance |
| 304 |
Durable behavioral agent contract |
Context management |
Located source instance |
| 305 |
Environment as instruction source of truth |
Skill authoring |
Located source instance |
| 306 |
Evidence-gated learning memory |
Context management |
Located source instance |
| 307 |
Compatibility-preserving migration tickets |
Workflow control |
Located source instance |
| 308 |
Explicit partial and invalid fixture intent |
Evaluation & feedback |
Located source instance |
| 309 |
Dependency-ready integration frontier |
Agent orchestration |
Located source instance |
| 310 |
Mechanics to checks, judgment to review |
Evaluation & feedback |
Located source instance |
| 311 |
Consequence-aware change evidence |
Output contracts |
Located source instance |
| 312 |
Question-shaped throwaway experiment |
Workflow control |
Located source instance |
| 313 |
Dependency-ready question rounds |
Workflow control |
Located source instance |
| 314 |
Question the knowledge owner through a send brief |
Workflow control |
Located source instance |
| 315 |
Co-locate concept rules and caveats |
Context management |
Located source instance |
| 316 |
Concept-scoped rejection memory |
Context management |
Located source instance |
| 317 |
Repair the missing premise before shortening |
Output contracts |
Located source instance |
| 318 |
Reread shared artifacts before incremental writes |
Workflow control |
Located source instance |
| 319 |
Dependency-calibrated verification adapters |
Tool use |
Located source instance |
| 320 |
Compare structural variants in shared context |
Evaluation & feedback |
Located source instance |
| 321 |
Differentiated design briefs |
Agent orchestration |
Located source instance |
| 322 |
Keep tool roots separate from project roots |
Tool use |
Located source instance |
| 323 |
Persist a master design system with page-specific overrides |
Workflow control |
Located source instance |
| 324 |
Verify a retrieved UI recommendation before reuse |
Workflow control |
Located source instance |
| 325 |
Check privacy and evidence usefulness separately |
Safety & trust |
Located source instance |
| 326 |
Record reversible decisions with their rationale and cost |
Workflow control |
Located source instance |
| 327 |
Preserve linkage with stable redaction placeholders |
Safety & trust |
Located source instance |
| 328 |
Qualify observations by when they were captured |
Evaluation & feedback |
Located source instance |
| 329 |
Probe-to-Saturation Interview |
Workflow control |
Located source instance |
| 330 |
Same-Artifact All-Reviewer Gate |
Agent orchestration |
Located source instance |
| 331 |
Atomic Budget Admission Before Enqueue |
Tool use |
Located source instance |
| 332 |
Audience, Participation and Delivery Separation |
Safety & trust |
Located source instance |
| 333 |
Bind Behavioral Specs to Enforcement and Tests |
Context management |
Located source instance |
| 334 |
Hard Blockers Override Aggregate Scores |
Evaluation & feedback |
Located source instance |
| 335 |
Conditioned Description Is Not Independent Validation |
Evaluation & feedback |
Located source instance |
| 336 |
Cross-Skill Evidence Distillation |
Skill authoring |
Located source instance |
| 337 |
Decision-Enabling Audit Reasons |
Skill authoring |
Located source instance |
| 338 |
Deterministic Parse with Uncertain-Case Escalation |
Tool use |
Located source instance |
| 339 |
One Canonical Owner per Documentation Fact |
Context management |
Located source instance |
| 340 |
Drain Completion before Stopping the Observer |
Workflow control |
Located source instance |
| 341 |
Eligibility Before Expensive Scoring |
Workflow control |
Located source instance |
| 342 |
Evidence-Gated Rule Promotion |
Skill authoring |
Located source instance |
| 343 |
Promote the Exact Verified Artifact |
Workflow control |
Located source instance |
| 344 |
Choose Metrics from Failure Costs |
Evaluation & feedback |
Located source instance |
| 345 |
Golden-File Convention Pointers |
Context management |
Located source instance |
| 346 |
Exact-Draft Epoch-Bound Approval |
Safety & trust |
Located source instance |
| 347 |
Identifiable Calibration with a Holdout |
Tool use |
Located source instance |
| 348 |
Empty-vs-Incomplete Recall |
Context management |
Located source instance |
| 349 |
Keep the Stop Supervisor Outside the Worker |
Agent orchestration |
Located source instance |
| 350 |
Gap-Driven Iterative Retrieval |
Context management |
Located source instance |
| 351 |
Preserve Behavioral Control Semantics Across Localization |
Skill authoring |
Located source instance |
| 352 |
Measurement Unit and Scope Boundaries |
Evaluation & feedback |
Located source instance |
| 353 |
Recall Freshness and Supersession |
Context management |
Located source instance |
| 354 |
Related-Input Invariant Oracle |
Evaluation & feedback |
Located source instance |
| 355 |
Ordered Actions and Final-State Audit |
Tool use |
Located source instance |
| 356 |
Observe a Blocked Worker from a Separate Control Path |
Agent orchestration |
Located source instance |
| 357 |
Compare Candidate and Baseline on the Same Trial Panel |
Evaluation & feedback |
Located source instance |
| 358 |
Point-in-Time Context Admission |
Context management |
Located source instance |
| 359 |
Propagate Failed Prerequisites |
Evaluation & feedback |
Located source instance |
| 360 |
Raw Evidence and Synthesis Separation |
Context management |
Located source instance |
| 361 |
Retry Success vs Repeated Reliability |
Evaluation & feedback |
Located source instance |
| 362 |
Adopt-Extend-Compose-Build Decision |
Workflow control |
Located source instance |
| 363 |
Evidence Bound to the Same Stable Bytes |
Tool use |
Located source instance |
| 364 |
Scope-Bound Evidence Memory |
Context management |
Located source instance |
| 365 |
Exclude Self-Generated Feedback Events |
Workflow control |
Located source instance |
| 366 |
Process Success and Semantic Completion Handshake |
Tool use |
Located source instance |
| 367 |
Aggregate and Slice Evaluation Gates |
Evaluation & feedback |
Located source instance |
| 368 |
Reusable Source-Derived Voice Profile |
Prompt design |
Located source instance |
| 369 |
Structured State and Rendered Evidence |
Tool use |
Located source instance |
| 370 |
Technical Facts vs Product Constraints |
Prompt design |
Located source instance |
| 371 |
First Verifiable Vertical Slice |
Workflow control |
Located source instance |
| 372 |
Validate the Proposed Mutation |
Tool use |
Located source instance |
| 373 |
Preserve Literal Examples Across Invocation Templates |
Skill authoring |
Located source instance |
| 374 |
Preserve Executable Identifiers Across Localization |
Skill authoring |
Located source instance |
| 375 |
Isolate Optional Capability Loading Failures |
Tool use |
Located source instance |
| 376 |
Version-Grounded Documentation Retrieval |
Context management |
Located source instance |
| 377 |
Accepted-obligation amendment ledger |
Context management |
Located source instance |
| 378 |
Separate behavior success from intervention benefit |
Evaluation & feedback |
Located source instance |
| 379 |
Capability-tiered instruction digest |
Context management |
Located source instance |
| 380 |
Comparable-coverage trend reporting |
Evaluation & feedback |
Located source instance |
| 381 |
Budget context by when it loads |
Context management |
Located source instance |
| 382 |
Keep claims within the observed channel |
Evaluation & feedback |
Located source instance |
| 383 |
Revoke the prior grant when narrowing access |
Safety & trust |
Located source instance |
| 384 |
Independent completion witness |
Evaluation & feedback |
Located source instance |
| 385 |
Review both overlapping completion orders |
Evaluation & feedback |
Located source instance |
| 386 |
Owner-token terminal cleanup |
Workflow control |
Located source instance |
| 387 |
Persona-grounded peer benchmarking |
Evaluation & feedback |
Located source instance |
| 388 |
Predeclare trials and verdict rules |
Evaluation & feedback |
Located source instance |
| 389 |
Codify a proven run with replay |
Skill authoring |
Located source instance |
| 390 |
Verify a save against the selected target |
Context management |
Located source instance |