Productivity
smart-sourcing
Selects optimal sources for tool calls, balancing accuracy with token cost. Use before research tasks or when deciding whether a claim needs verification
Engineering
math-review
Verifies math-heavy code for algorithmic correctness and numerical stability
Engineering
safety-critical-patterns
Applies NASA Power of 10 rules for safety-critical verifiable code
Productivity
pdf-proof
Use this skill when the user needs visual proof that specific values exist in a PDF — not just to read a PDF, but to see exactly where a number, amount, clause, or field appears on…
Research
tribunal
Adversarial tribunal that exploits agent sycophancy for high-fidelity analysis. Three agents — a maximalist finder, an adversarial challenger, and a referee — compete under asymmet…
Testing
go
Opens the running app in a browser and verifies recent UI changes actually work. Use whenever the user wants a quick smoke test or sanity check of recent work, or says "go", "open …
Productivity
verify-prd
Initiative-level PRD acceptance gate. Resolves a PRD reference, reads its body and child work hierarchy, and halts if any top-level item is incomplete. On success, runs spec confor…
Research
eac-habre-jeanne-sulzer
Verification-first methodology for the Extraordinary African Chambers (Habré trial, Dakar). Verifies citations via forumchambresafricaines.org and legal-tools.org; covers the 30 Ma…
Productivity
fable-advisor
Consults a premium reasoning model for high-stakes choices—architecture, schemas, API contracts, workflows, vendor selection, migrations—before irreversible commitments. Reviews un…
Business
bates-citation-verification
Citation format and verification protocol for legal documents referencing Bates-stamped source records. Defines bracket citation format, multi-source citation rules, quotation hand…
Testing
parity-replace
Performs parity-based replacement: partitions a feature into page-level phases, fills locator-mapping gaps, and runs adversarial review on uncommitted diffs. Reads current code as …
Testing
dashboard-e2e-testing
Verify a dashboard or agent-facing surface actually works, not just that its components render. Use before treating any dashboard/backend change as done, when adding a new surface …
Documentation
forge-citation
Citation discipline for AI-generated outputs. Defines when to cite, what qualifies as verifiable, URL+quote+file-path verification, inference vs fact marking, and what to avoid. Pr…
Testing
manual-test-planning
Produces plain-language manual test plans from supplied context: an executive summary, named test list, and step-by-step verification details with expected outcomes. Organizes cont…
Data
product-search
Performs ingredient-level verification of supplement and product recommendations by cross-checking against user constraints in MEMORY.md and medical-history.md, fetching actual pro…
Productivity
achieve
Tracks a long-running objective as an auditable goal artifact under charness-artifacts/goals/, surfacing slice progress and verification throughout, then closes with honest non-cla…
Research
scsl-rscsl-special-court-for-sierra-leone-jeanne-sulzer
Verification-first research skill for the Special Court for Sierra Leone and its Residual Court, cross-checking citations against rscsl.org and legal-tools.org. Covers Taylor, AFRC…
Engineering
pie-dimensional-analysis
Performs mathematical verification for physical calculations: unit-tracking algebra, PhysicalQuantity patterns for compound units, SI/Imperial handling, Buckingham pi theorem, and …
Testing
loop-evals
Designs the evaluation harness for agent loops, establishing trustworthy verification through a 7-layer suite with false-completion-rate and repair-productivity as core metrics. En…
General
conscientiousness
Apply thoroughness and diligence in execution through systematic checking, completeness verification, follow-through on commitments, and disciplined task completion. Use before mar…
Design
designer-frontend-contract
Universal contract between designer and frontend: designer freezes approved mockup into oracle (manifest + baseline renders + inventory + tokens + acceptance criteria); frontend po…
Productivity
ultraloop
Completion discipline that treats a task as done only when the result is verified and demonstrably works. Inspects the repo first, sets explicit assumptions and a Definition of Don…
Automation
spec-driven-migration
Performs reference-sweep migrations by replacing `from_pattern` with `to_pattern` across large codebases using four-layer verification: classify-first, dual-path identical edits, f…
Engineering
k8s-fix
Executes approved Kubernetes fixes through a bounded plan-review loop with adversarial checks, red-first tests, minimal edits on a fresh branch, parallel read-only reviews, synthes…
Engineering
repo-task-proof-loop
Repo-local workflow skill for large coding tasks. Initializes .agent/tasks/TASK_ID artifacts, installs project-scoped Codex and Claude subagents, updates AGENTS.md plus the repo's …
AI / ML
bedrock-automated-reasoning
Amazon Bedrock Automated Reasoning performs mathematical verification of AI outputs against formal policy rules, reaching up to 99% accuracy. Ideal for validating healthcare protoc…
Engineering
deliver-software-project
Autonomously delivers a bounded software project or milestone from repository inspection through implementation, repair, verification, and evidence-based handoff under a one-time a…
Testing
qa-gen
Generates a test plan from spec criteria, classifying each by verification method and writing 06_qa.md with scoped, traced scenarios. UI criteria map to browser runs; backend crite…
Documentation
scoring-and-release-gate
Scores documentation packages against a ten-dimension rubric with cited justifications, then evaluates eighteen release conditions to output a binary ready/conditional/not-ready ve…
Engineering
zero-hallucination-coder
Delivers a strict Discuss-Map-Decompose-Execute-Verify workflow that anchors every line of code to verified structure, eliminating invented APIs, assumed imports, and placeholder l…
Research
nuremberg-tokyo-jeanne-sulzer
Verification-first research aid for post-WWII tribunals: IMT Nuremberg, twelve NMTs, and IMTFE Tokyo. Verifies citations against official records (Blue/Green Series, Pritchard-Zaid…
DevOps
deploy-verify
Post-deploy verification workflow that runs smoke tests across observability, database, tracing, and browser checks. Detects context automatically, surfaces new errors with root-ca…
Research
source-locked-verification
Enforces source-locked verification: answers only from user-provided materials or verified online sources, with every factual claim anchored to a cited reference. Use for document …
Testing
user-uat
Executes a provided UAT or manual-smoke plan by running each step, capturing output, and auto-verifying mechanical checks (exit codes, output matches, refusal text, DB/HTTP/file/lo…
Research
eccc-jeanne-sulzer
Verification-first research aid for the ECCC (Khmer Rouge Tribunal). Every citation is validated against eccc.gov.kh, the ECCC archive, or legal-tools.org. Covers Case File structu…
Productivity
brain-link-discipline
Enforces that every brain page deliverable includes a working link in the same message. Derives paths mechanically, pushes before linking, verifies hosted remotes, and applies a fa…
Testing
in-fact-check
Verifies Indiana legal citations across four passes: checks rules, statutes, and cases against authoritative sources; confirms internal consistency of names, numbers, and dates; al…
Engineering
dead-code-verification
Confirms whether a symbol, file, route, config path, or feature can be safely removed by checking static and dynamic references across registries, builds, templates, plugins, scrip…
Security
bsv-cx-verify
Verify any Bitcoin (BSV) Merkle proof — free, no account, trusting no one including bsv.cx — against a block-headers node bsv.cx syncs itself over the BSV P2P network. Send a raw T…
Research
witsoc-generator
Generates, repairs, and validates `.wit` proof artifacts for mathematics and rigorous arguments, including structural checks, context building, receipt tracking, and optional Lean …
Automation
claude-managed-agents-webhooks
Receive and verify managed agent webhooks. Use when setting up webhook handlers, debugging signature verification, or handling agent session and vault events such as session.status…
DevOps
post-deploy-verifier
Produces post-deploy verification plans that map each acceptance criterion to live signals (log queries, dashboards, smoke tests, flag checks, DB invariants) with negative checks a…
Research
paper-verify
Verifies reference existence, DOI validity, metadata accuracy, retraction status, and checks for overstated conclusions using live data sources rather than memory. Returns structur…
Engineering
devlyn:resolve
End-to-end autonomous pipeline for any engineering task—bug fixes, features, refactors, debugging, or PR reviews. Accepts free-form goals or formal specs, then plans, implements, g…
Engineering
neckbeard
Routes non-trivial change requests through framing, discovery, design, implementation, gated review, verification, and authorized release, composing specialist skills rather than r…
Research
cps-rca-central-african-republic-jeanne-sulzer
Verification-first methodology for the Special Criminal Court in the Central African Republic. Validates citations against JusticeInfo, FIDH, HRW, and legal-tools.org sources. Cove…
Testing
dos-witness-claim
Route subagent claims through a witness step that extracts verifiable effects (git commits, files, DB rows) and confirms them via independent read-back before folding. Driven by do…
General
false-completion-gate
Fires before any state-changing task claims Done, decomposing completion statements into verifiable assertions then auditing them against the session's tool-call ledger and fresh a…
Engineering
semi-formal-review
Semi-formal review that issues a CORRECT / LIKELY_CORRECT / CONCERNS / BUGGY verdict on whether a change fully meets its spec without introducing regressions. Mandatory post-implem…
Testing
aif-verify
Verify completed implementation against the plan. Checks that all tasks were fully implemented, nothing was forgotten, code compiles, tests pass, and quality standards are met. Use…
Testing
jsgui3-verify
Verifies that jsgui3 SSR and client-activated apps render and behave correctly, exposing common failure modes where the UI appears functional but is not. Invoke before any src/ui/s…
Research
mandatory-verification
Mandatory external verification workflow required for all non-trivial factual claims before asserting them as true. Triggers on research, factual queries, current data, statistics,…
AI / ML
verified-capability-evolver
Adds verification, rollback, and gated promotion to capability evolution. Before promoting any learning, establishes a deterministic spec, validates output through SettlementWitnes…
Research
icalp-experiments
Determines whether computational support (SAT/SMT verification, exhaustive enumeration, or case analysis) can legitimately strengthen an ICALP theorem's proof without converting th…
Research
Intrinsic Institute
This repository implements a comprehensive, systematic verification protocol ensuring that every algorithmic implementation, computational construct, and symbolic representation co…
Testing
delivery-record-verify
Validates Delivery Record files against the kanopi/delivery-record schema. Resolves the schema version from predicate_type, enforces required fields, validates check values (pass|f…
Engineering
formal-methods-and-model-checking
Provides mathematical proof that a system satisfies safety properties via model checking and temporal logic. Activates on mentions of TLA+, TLC, Alloy, invariants, counterexamples,…
Testing
aiw-verification
Enforces verification of every change before marking work complete. Requires explicit justification of checks performed, evidence hierarchy from static analysis to full end-to-end …
Testing
counterexample-explainer
Diagnoses why counterexamples violate specs by tracing state changes across formal models, requirements, and tests—comparing expected vs actual behavior, pinpointing root causes, a…
Testing
landing-check
Verifies that changes actually landed on disk by reading ARTIFACTS (diffs, records, reports) rather than session memory—confirms execution paths, validates stated claims, and flags…
Showing the top 60 of 1,731. See the full list →