Skill Claude CodeCodex
Use to evaluate architecture, design proposals, refactors, module boundaries, dependency direction, data flow, concurrency model, and maintainability tradeoffs across any codebase.
Skill Claude CodeCodex
Use to evaluate architecture, design proposals, refactors, module boundaries, dependency direction, data flow, concurrency model, and maintainability tradeoffs across any codebase.
Skill Claude CodeCodex
Use for structured research on an unfamiliar or complex codebase before planning, reviewing, or editing. Delegates mapping, ownership, dependency, test, and risk discovery to focused subagents and returns an evidence-led research brief.
Skill Claude CodeCodex
Use when implementation or review depends on external documentation, standards, protocols, APIs, SDKs, changelogs, or version-specific behavior. Produces a concise source dossier and implementation implications.
Skill Claude CodeCodex
Use for rigorous pull request, branch, patch, or diff review. Runs focused review subagents for correctness, security/reliability, tests, docs/API behavior, and release risk, then synthesizes actionable findings.
Skill Claude CodeCodex
Use to design, run, or review performance work with reproducible A/B baselines. Delegates benchmark inventory, profiling, correctness guardrails, and result interpretation to subagents.
Skill Claude CodeCodex
Use before merging, shipping, publishing, or handing off work. Runs subagents for verification, regression risk, operational readiness, docs, compatibility, and unresolved follow-ups.
Skill Claude CodeCodex
Use to compare any codebase against a specification, API contract, protocol, policy, RFC, security baseline, or product requirements document. Builds a traceability matrix and prioritizes compliance gaps.
Agent Cursor
Read-only repository explorer for architecture, execution paths, ownership, dependencies, and validation surfaces. Use before planning, reviews, or edits.
Agent Cursor
Read-only reviewer for functional correctness, behavior regressions, edge cases, compatibility, data model risks, and missing tests.
Agent Cursor
Use for CSV workbench (/csv) workflows — profiling, merge recipes, session cleanup. You are an external Cursor development agent; you work from the repo/worktree and use Open-FDD MCP or REST. You are not part of the Open-FDD edge runtime and are not embedded in the dashboard.
Agent Cursor
Documentation and source researcher for external APIs, standards, dependencies, SDKs, protocols, release notes, and version-specific behavior.
Agent Cursor
Use for Model (/model) and SQL FDD (/sql-fdd) workflows. You are an external Cursor development agent invoked outside the Open-FDD dashboard; use MCP/REST from the repo/worktree. Not part of the GHCR edge runtime.
Agent Cursor
Use for complex Open-FDD work spanning docs, Rust edge, scripts, deploy safety, auth, or multi-service validation. Cursor is an external operator agent — not embedded in the Open-FDD dashboard.
Agent Cursor
Benchmark and profiling specialist for baseline discovery, A/B comparison, latency, throughput, memory, allocations, and workload validity.
Agent Cursor
Read-only release and operational risk reviewer for rollout, migration, compatibility, observability, docs, rollback, packaging, and support readiness.
Agent Cursor
Read-only reviewer for security, privacy, secrets, authz/authn, input validation, concurrency, resource exhaustion, error handling, and fail-closed behavior.
Agent Cursor
Use for simple failures where a worker/mini model is sufficient. External Cursor agent — not part of Open-FDD edge runtime.
Agent Cursor
Requirement tracer for specs, API contracts, protocols, policy docs, acceptance criteria, and compliance matrices.
Agent Cursor
Verification specialist for tests, builds, lint, typecheck, CI gates, coverage gaps, flaky tests, and independent validation of completed work.
Cursor rule Cursor
Never demo stale GHCR web; pick images by OCI created; gate Caddy URLs.
Cursor rule Cursor
Phase 1 React/Rust parity — require openfddagentspec + React SPA skill.
Instructions file CodexOpenCode
AGENTS.md instructions for bbartling/open-fdd, covering agent guide (container stack + external agents), start session, or: react (no fieldbus) / csv, rust hygiene (agents) and safe scripts.
Skill Claude CodeCodex
Use when enforcing Open-FDD product boundaries: production DataFusion SQL vs pandas oracle, vibe19/vibe20 ownership, dual cookbooks, edge/os never-delete, ownership.yaml, forbidden imports. Triggers on: architecture, ownership, pandas fallback, cookbook delete, Milestone A Phase 0.
Skill Claude CodeCodex
Use when editing rule cookbooks, parity matrix, or cookbook CI (cookbook-parity.yml, cookbookparitycheck.py). Triggers on: pandas cookbook, SQL cookbook, parity-matrix, gap-matrix, rule headings.