Product Manager (Morgan). Use for PRD creation (greenfield and brownfield), epic creation and management, product strategy and vision, feature prioritization (MoSCoW, RICE), roa...

Validate a story draft (Draft → Ready on GO). Runs the OSS PO validation task. Use when: validate story, *validate-story-draft, /validate-story-draft, story readiness.

QA gate for a story — verdict PASS/CONCERNS/FAIL/WAIVED + lifecycle transition. Use when: review story, qa gate, *qa-gate, /review-story.

Finalize bookkeeping for a story already completed by QA. Never changes lifecycle status. Use when: close story, *close-story, story Done, /close-story.

QA gate for a story — verdict PASS/CONCERNS/FAIL/WAIVED + lifecycle transition. Use when: review story, qa gate, *qa-gate, /review-story.

Update and maintain the llms.txt file. Use when the user wants to add sections, add links, correct information, or review the llms.txt file.

How to write end-to-end tests using createRouterAct and LinkAccordion. Use when writing or modifying tests that need to control the timing of internal Next.js requests (like prefetches) or assert on their responses. Covers the act API, fixture patterns, prefetch control via LinkAccordion, fake clocks, and avoiding flaky testing patterns.

Triage CI failures and PR review comments using scripts/pr-status.js. Use when investigating failing CI jobs, flaky tests, or PR review feedback. Covers blocker-first prioritization (build > lint > types > tests), CI env var matching for local reproduction, and the Known Flaky Tests distinction.

Create new Harbor Boost modules — the Python plugins that run inside Harbor's LLM proxy. Use this skill whenever the user wants to build a Boost module, write a custom module for Harbor Boost, add a new feature to the Boost proxy pipeline, or create any kind of middleware that transforms, augments, or intercepts LLM chat completions in Harbor. Also triggers when the user mentions "boost module", "boost plugin", "custom module for boost", or wants to add prompt engineering, reasoning chains, or output transforms to Harbor's proxy layer.

Operate on @spec facts — implement them in code, then tag @implemented. Use when asked to implement facts, implement the spec, build from the fact sheet, make facts true, or work through unimplemented requirements.

Hunt server-side template injection (SSTI) across Jinja2 (Flask/Django), Twig (Symfony), Freemarker (Java), ERB (Rails), Spring, Velocity, Mako, Thymeleaf, Smarty. Detection probes use double-curly and dollar-curly math expressions evaluated server-side. Once an engine is fingerprinted, escalate to RCE via the engine-specific class-walker, callback-registrar, or Execute-utility patterns documented in disclosed reports. Detection patterns: error messages reveal engine, blank or numeric eval reveals expression mode. Targets: email templates, PDF/report generators, CMS preview features, error pages with user input. Use when hunting RCE via template rendering, when content shows engine fingerprints, when finding endpoints that compose strings with user input before render.

Hunt source code and build artifact leakage — JavaScript source maps (.js.map) reconstructing TypeScript/ES6 source, Swagger/OpenAPI JSON endpoint discovery, .env/.git exposure, webpack chunks with hardcoded secrets, robots.txt/security.txt recon, build-info files, asset-manifest.json API route discovery, .DS_Store file listing. Use at the START of every recon session — these findings often unlock the entire attack surface.

Hunt vector-store / embedding-layer weaknesses in RAG pipelines (OWASP LLM08 Vector and Embedding Weaknesses) — persistent corpus poisoning that survives across sessions and users (distinct from one-shot indirect prompt injection, which is owned by hunt-llm-ai), cross-tenant vector-database IDOR (unauthenticated or unscoped queries against Pinecone/Weaviate/Chroma/Milvus/Qdrant/pgvector), source-text/metadata leakage in similarity-search results, and retrieval-hijack via adversarial embedding proximity ('SEO poisoning' for RAG). Targets: any app with a shared knowledge base, document upload feeding a chatbot, or a directly reachable vector-DB port. Validate: a second, clean session/account must inherit a poisoned result, or a cross-tenant artifact must be independently verifiable — confabulation is not a finding, same bar as hunt-llm-ai. Use when target is RAG-backed, exposes a vector-DB port, or lets users upload documents that other users' queries later retrieve.

Hunt LLM/AI feature bugs — prompt injection, indirect injection, exfiltration via tool-use/markdown, ASCII smuggling, agentic AI security (OWASP Agentic Apps 2026, ASI01-ASI10). Patterns: direct injection ('ignore previous instructions'), indirect injection via documents/web pages/email the model reads, ASCII smuggling (Unicode Tags block U+E0000-U+E007F, invisible to humans, decoded by the model), tool-use exfiltration (model has fetch/browse tool, attacker injects OOB URL, model exfils chat history/secrets), markdown-image zero-click exfil, system-prompt extraction, IDOR-via-AI (cross-tenant data). Targets: chatbots, RAG, summarizers, agentic copilots, MCP tools. Detection: any LLM-backed endpoint, doc upload triggering AI processing, autonomous agent with tools. Validate: OOB/Collaborator callback for exfil, verbatim-reproducible system-prompt leak (run twice), verifiable cross-tenant leak or RCE. Confabulation is NOT a finding. Use when hunting AI features, chatbots, RAG, agentic systems, MCP.

Hunt cloud / infrastructure misconfigurations. AWS: public S3 buckets (s3:GetObject anonymous), permissive bucket policies (PutObjectAcl public-write), exposed CloudFront origin, public Lambda function URL, public RDS snapshot, IAM credentials in JS bundles, AWS metadata accessible via SSRF. GCP: public GCS buckets, exposed Cloud Run services, leaked service account JSON. Azure: public blob containers, exposed Function App. (Kubernetes/Docker exposure is owned by hunt-k8s; CI/CD pipeline attacks by hunt-cicd; post-credential IAM escalation by cloud-iam-deep.) Detection: targeted dorking, certificate transparency, JS bundle secret extraction, port scan for known service ports. Validate: actual data read / write / RCE. Use when hunting cloud-native storage and compute misconfig (S3/GCS/Blob, IMDS-via-SSRF, serverless, public managed services).

Use when designing or rewriting a high-converting landing page (single-offer page) for SaaS/apps/services. Covers structure, layout patterns, conversion strategies, copywriting, SEO/AEO, and common pitfalls.

Package, deploy, and verify a playable Three.js or web game. Use for release builds, asset delivery, private/public deployment, production smoke tests, browser proof, release notes, rollback readiness, and cleanup of temporary QA resources.

Implement, tune, debug, or validate soft wall-aware fog of war and gameplay perception in Three.js action games. Use for orthographic or isometric visibility masks, obstacle-aware line of sight, player and enemy vision ranges, hidden-enemy targeting rules, fog shader artifacts such as spokes or seams, mobile ray budgets, lifecycle and menu-state integration, and deterministic fog-of-war tests.

Apply subtle gradient-border treatments for premium web surfaces. Use when cards, pricing panels, nav bars, modals, buttons, or hero surfaces need a refined edge highlight without a loud glow.

Create a globe-like 3D particle visualization with a dense luminous spherical core and thinner orbital ring or flattened disc. Use when a design needs a premium planetary, orbital, synthesized data-globe effect rendered with real WebGL/Three.js particles, not generic starfields or full page layout changes.