Blind Probe Author — Development agent for Claude Code
Writes efficacy-eval scenarios, retrieval probes, and held-out criteria for a document type WITHOUT reading this library's templates, guides, rubrics or manifest.
How to install Blind Probe Author
Installs to ~/.claude/agents/product-on-purpose-product-lifecycle-templates-blind-probe-author.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/product-on-purpose/product-lifecycle-templates/HEAD/.claude/agents/blind-probe-author.md -o ~/.claude/agents/product-on-purpose-product-lifecycle-templates-blind-probe-author.md Restart Claude Code, or start a new session, for it to be picked up.
What Blind Probe Author does
name: blind-probe-author description: Writes efficacy-eval scenarios, retrieval probes, and held-out criteria for a document type WITHOUT reading this library's templates, guides, rubrics or manifest. Use when the eval protocol's section 6 blinding rule applies, which is any time a scenario, probe or held-out criterion is authored or rewritten. Maintainer-internal; not shipped to library users. tools: WebSearch, WebFetch model: sonnet
You author evaluation material for a template librar
Alternatives in Development
- Context Restore — Expert Context Restoration Specialist focused on intelligent, semantic-aware context retrieval and r... - wsho 31.9k ★
- Seam Analyzer — Hunts for a missing type at a seam — structure flattened and rebuilt downstream, hand-maintained lists held to 2.2k ★
- RAG Specialist — GAIA RAG and agentic-retrieval specialist 1.5k ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Manual Rule Liveness Prober
Probes a manifest manual rule for liveness via with/without-rule fresh-subagent dispatch, capturing a RED→GREE
What If
Opus 5 plan prober that walks a handed plan against a facet checklist and returns ranked blind-spot probes, ea
Dispatch Input Checker
Cold dispatch-input reality-check at the aif-dispatch station boundary. Given ONLY the dispatch input (kickoff
Typesafe Adversary
Red-teams TypeSafe integrations before users do. Probes the nine documented Jev failure modes, tests prompt in
Eval Skew Reviewer
Reviews diffs that add or change a write-time enrichment feature / retrieval channel for training-serving skew
Permit Extractor
Retrieval-only pass over OCRed permit documents. Given a section's questions and document paths, produces the
Related Skills
Fable Eval
Run the fable-mode eval suite (probes → pairwise judge → report). Costs tokens — runs headless claude many tim
Dotnet Refactor
Phase-gated refactoring/design loop - session-blind subagents build the map and traces, empirical probes prece
Claude Interaction Tests
Test whether your CLAUDE.md actually works. A Claude Code skill that restructures agent-facing docs (CLAUDE.md