verify-before-claim
claudeCheck claims against evidence; optionally pin executable acceptance criteria and obtain independent verification for work that needs a formal contract.
4 skills · 2 commands · hooks
Design and audit an agent system's eval ladder — the cheapest rung that catches each regression, the blind spot beside every green, judges validated by TPR/TNR, pass^k for irreversible actions. Use when designing or auditing a test/eval strategy.
Plugin eval-ladder from marketplace jrichlen in JRichlen/agent-plugins. Commands read from the repo's manifest — 13 days ago.
eval-ladderDesign and audit the eval ladder for an agent system: build tiers bottom-up from observed failures, pick the cheapest rung that can catch a given regression, name what each rung structurally cannot prove, validate every LLM judge against human labels (TPR and TNR separately, never raw agreement), an
/eval-ladder:eval-ladderDesign or audit an eval strategy — map coverage onto the surface × rung matrix, name each tier's blind spot, check every judge is validated, and rank the gaps.
Check claims against evidence; optionally pin executable acceptance criteria and obtain independent verification for work that needs a formal contract.
4 skills · 2 commands · hooks
Route every response element to exactly one voice: prose to human-voice, machine-read artifacts to machine-voice; plus ai-writing-mistakes (wording pass) and second-opinion (offer-only validation).
4 skills · 1 command · hooks
Compile deterministic, content-hashed agents from small behavior modules: a typed AgentQuery in, an immutable AgentImage and rendered harness-native subagent out; fails closed on conflicts.
1 skill · 2 commands · 1 MCP server · hooks
Keep a scannable developer journal that writes itself: discover the day's work from sessions, commits and prompts, interview the user, record a dated entry. Ships /dev-diary and /dev-diary-review.
2 skills · 2 commands
Classify a candidate action as PATCH/MINOR/MAJOR (semver-style blast-radius test) before acting, and rehearse the rollback of any irreversible one — name the restore path and exercise it. Use mid-task when unsure how much autonomy to take.
2 skills · 2 commands
Set up and troubleshoot secretless GitHub Actions -> Tailscale auth via Workload Identity Federation (WIF): GitHub OIDC exchanged for short-lived Tailscale tokens, with no stored API keys or OAuth secrets.
1 skill · 1 command · 1 agent
Create new skills, improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, update or optimize an existing skill, run evals to test a skill, or benchmark skill performance with variance analysis.
1 skill
Eval-gated LLM fine-tuning lifecycle (LoRA/QLoRA, DPO, GRPO/RLVR, vision SFT, quantized export). Use when fine-tuning an open-weights model end-to-end: building the eval harness first, selecting method and model, preparing data, training with Unsloth or TRL, gating checkpoints, and exporting GGUF/FP8.
10 skills · 2 commands · 3 agents
45 advanced engineering skills: agent designer, agent workflow designer, RAG architect, database designer + schema designer + SQL assistant, migration architect, observability designer, dependency auditor, changelog generator, API design reviewer, API test suite builder, CI/CD pipeline builder, MCP server builder, skill security auditor, skill tester, performance profiler, focused-fix, browser-automation, full-page-screenshot, git-worktree-manager, monorepo-navigator, codebase-onboarding, interview-system-designer, runbook-generator, spec-driven-workflow, secrets-vault-manager, env-secrets-manager, pr-review-expert, self-eval, tc-tracker, feature-flags-architect, kubernetes-operator, chaos-engineering, ship-gate, slo-architect (error-budget + multi-window burn-rate alerts per Google SRE Workbook), tech-debt-tracker, agent-regression-evals, coding-agent-rollout, commit-craft, agent-run-forensics and i18n-localization. Agent skill and plugin for Claude Code, Codex, Gemini CLI, Cursor, OpenClaw.
48 skills · hooks
Connect providers and targets, write eval suites, and set up or run red team scans with promptfoo
4 skills
Connect to your Phoenix instance to debug, evaluate, and improve LLM applications.
1 MCP server
Autonomous improvement engine. 14 commands with bounded defaults, chain handoff, and adaptive evals checkpoints. 95% token reduction via modular architecture.
1 skill · 14 commands · hooks
This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.