mishrasanjeev

Use this skill for any CaseOps bug triage, verification, reopen analysis, or review of another agent's bug-fix claim. Enforces fail-closed bug handling: reproduce, root cause, adjacent-path audit, regression coverage, strongest verification, and honest verdicts.

Use this skill for any CaseOps feature planning, implementation, product-gap analysis, module review, UX redesign, or PRD update. Forces every task to map to the unified PRD, current repo truth, source-data rules, user stories, and test IDs before work begins.

Use this skill whenever the user asks to ingest, embed, populate, backfill, or improve quality of the CaseOps authority corpus (Supreme Court / High Court judgments) in Postgres/pgvector on GCP. Also use when the user asks for a retrieval quality probe or a 0-5 rating on the embedded vectors. Enforces the correct per-bucket pipeline (ingest → Layer-2 metadata → title-chunk embed → quality probe) that prevents the "placeholder title poisons embeddings" failure mode.

Use this skill for any CaseOps enterprise-readiness audit, scale-hardening review, architecture or security gap review, or any scan of docs/WORK_TO_BE_DONE.md. Enforces fail-closed gap classification, code-vs-doc proof, and durable ledger updates.

Use this skill for whole-repo quality audits, release-readiness reviews, exhaustive test-matrix planning, security reviews, or requests to make CaseOps quality gates stricter. Forces evidence-backed inventory, route/page/test matrices, verification results, and durable audit docs.

Create distinctive, production-grade frontend interfaces with high design quality. Generates creative, polished code that avoids generic AI aesthetics. Use when the user asks to build web components, pages, artifacts, posters, or applications, or when any design skill requires project context. Call with 'craft' for shape-then-build, 'teach' for design context setup, or 'extract' to pull reusable components and tokens into the design system.

Enterprise implementation guardrails for Grantex Commerce V1 across grantex.dev and agenticorg.ai. Use when reviewing, planning, implementing, testing, or signing off Grantex Commerce, agentic commerce, merchant AI-native onboarding, MCP/UCP/ACP publishing, merchant catalog/inventory/pricing, Commerce Passports, Plural payments, webhooks, APIs, database models, UI/UX, retries, observability, release gates, or production hardening from GRANTEX_COMMERCE_V1_BUILD_SPEC.md.

Post-merge verification pass for Grantex. After any PR merges to main or ships to prod, run this pass without being asked — verify README, Mintlify docs, landing page, SDK READMEs (TS, Python, Go), integration docs, and SEO surfaces are consistent with what just shipped. Use this skill whenever a PR merges, a release goes out, or a feature ships.

Expert review guidelines for auditing repositories, pull requests, and system designs. Use when performing code review to find functional gaps, security risks, correctness bugs, maintainability problems, missing tests, and architectural weaknesses without making direct code changes.

Use this skill for enterprise hardening work in the AgenticOrg repository: whole-codebase audits, release sign-off, architecture review, and any meaningful code change that must meet enterprise-grade standards. Apply it for backend API changes, auth or tenant isolation work, billing, connectors, secrets, workflows, migrations, async runtime behavior, frontend auth flows, worker or deployment changes, and any task where correctness, security, scale, and production safety matter more than speed. This skill is especially important when touching FastAPI routes, request authorization, multi-tenant data access, Redis or Celery, workflow orchestration, Alembic migrations, health checks, metrics, or React session handling.

Use this skill for every bug fix, reopen analysis, QA bug-sheet review, regression triage, and release sign-off in the AgenticOrg repository. It forces fail-closed verdicts, tester-step replay, sibling-path sweeps, production-vs-main honesty, and explicit residual-risk reporting. Use it whenever the work product says Fixed, Partially closed, Not reproducible, Already fixed, Enhancement, Duplicate, or Release sign-off.