Lead Checker — Development agent for Claude Code
Pursues leads from the aligned pool: triages the candidate observations from scibraid observe, checks the few worth the effort against the sources and the wider literature, and records each as a lea.
How to install Lead Checker
Installs to ~/.claude/agents/phelps-sg-scibraid-lead-checker.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/phelps-sg/scibraid/HEAD/agents/lead-checker.md -o ~/.claude/agents/phelps-sg-scibraid-lead-checker.md Restart Claude Code, or start a new session, for it to be picked up.
What Lead Checker does
name: lead-checker description: Pursues leads from the aligned pool: triages the candidate observations from `scibraid observe`, checks the few worth the effort against the sources and the wider literature, and records each as a lead with its checks. Use from the pursue-leads skill. Give it any focus the user asked for (a topic, a subgraph, a particular lead to revisit); with none it triages the whole pool. model: fable tools: Bash, Read, Write, Edit, Grep, Glob, WebFetch, WebSearch skills:
Alternatives in Development
- Idea Agent — Literature search and hypothesis formation 728 ★
- Librarian Critic — Literature quality critic 250 ★
- Creational Auditor — Creational patterns auditor 63 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Ana Learn
Ana Learn — quality gardener. Triages findings, promotes rules, routes observations.
Thales Prober
Thales Explorer persona. Pursues a current branch deeper by staking one specific testable claim. Spawned by th
Observability Guardian
Use when validating C6 (telemetry-by-default) — 100% trace coverage on production runs, instrumentation block
Claude Code Guru
Use PROACTIVELY and automatically — do not wait to be asked — as the operator's authority on ALL things Claude
Ga4 Attribution
GA4 attribution analyst. Analyzes channel performance at each funnel step, attribution model and lookback wind
Cross Functional Liaison
Use proactively for meeting notes, stakeholder updates, decision logs, and cross-functional communication. Han
Related Skills
Think Idea Quality Audit
Decide which of a batch of candidate ideas are worth committing to, by scoring them on explicit quality dimens
Honeycomb
AI coding agents forget. They forget across sessions, and they forget across tools. A decision you reached in
Lea
Many Claude Code agents. None of your machine.