review-agent
25Agent Claude Code
Review implementation by comparing plan (intent) vs Braintrust session (reality) vs git diff (changes).
not rated 3.9k
+2 8mo ago A 25 tokens
original MIT
scout
26Agent Claude Code
Codebase exploration and pattern finding.
not rated 3.9k
+2 8mo ago A 9 tokens
original MIT
scribe
27Agent Claude Code
Documentation, handoffs, session summaries, and ledger management.
not rated 3.9k
+2 8mo ago A 13 tokens
original MIT
Agent Claude Code
Analyze Claude Code sessions using Braintrust logs.
not rated 3.9k
+2 8mo ago A 13 tokens
original MIT
sleuth
29Agent Claude Code
General bug investigation and root cause analysis.
not rated 3.9k
+2 8mo ago A 10 tokens
original MIT
spark
30Agent Claude Code
Lightweight fixes and quick tweaks.
not rated 3.9k
+2 8mo ago A 8 tokens
original MIT
surveyor
31Agent Claude Code
Migration and upgrade review.
not rated 3.9k
+2 8mo ago A 7 tokens
original MIT
Agent Claude Code needs its repo
Validate plan tech choices against current best practices and past precedent.
not rated 3.9k
+2 8mo ago A 14 tokens
original MIT
braintrust
33Plugin Claude Code
Bundles 1 skill · 32 tokens together
Enables AI agents to use Braintrust for LLM evaluation, logging, and observability. Provides correct API usage, working examples, and helper scripts for common operations.
not rated 3.9k
+2 8mo ago A
tokens not measured
original MIT
Skill Claude Code
Prevent agent output from polluting the main context window.
not rated 3.9k
+2 8mo ago B Socket: passSnyk: pass 8 tokens
original MIT
Skill Claude Code
When the user asks to implement something, use implementation agents to preserve main context.
not rated 3.9k
+2 8mo ago A Socket: passSnyk: passSkillSpector: pass 10 tokens
original MIT
Skill Claude Code
Standard multi-agent pipeline for implementation tasks.
not rated 3.9k
+2 8mo ago B Socket: passSnyk: pass 9 tokens
original MIT
Skill Claude Code needs its repo
Guide for integrating Agentica SDK with Claude Code CLI proxy.
not rated 3.9k
+2 8mo ago A Socket: passSnyk: passSkillSpector: warn 19 tokens
original MIT
Skill Claude Code
Reference guide for Agentica multi-agent infrastructure APIs.
not rated 3.9k
+2 8mo ago A Socket: passSnyk: passSkillSpector: pass 14 tokens
original MIT
Skill Claude Code
Write reliable prompts for Agentica/REPL agents that avoid LLM instruction ambiguity.
not rated 3.9k
+2 8mo ago B Socket: passSnyk: pass 21 tokens
original MIT
agentica-sdk
40Skill Claude Code
Build Python agents with Agentica SDK - @agentic decorator, spawn(), persistence, MCP integration.
not rated 3.9k
+2 8mo ago A Socket: passSnyk: passSkillSpector: warn 23 tokens
original MIT
Skill Claude Code needs its repo
Agentica server + Claude proxy setup - architecture, startup sequence, debugging.
not rated 3.9k
+2 8mo ago A Socket: passSnyk: passSkillSpector: warn 18 tokens
original MIT
Skill Claude Code
Spawn Agentica multi-agent patterns.
not rated 3.9k
+2 8mo ago A Socket: passSnyk: passSkillSpector: pass 11 tokens
original MIT
Skill Claude Code
When working with Agentica's async REPL harness for testing.
not rated 3.9k
+2 8mo ago A Socket: passSnyk: pass 10 tokens
original MIT
leann-search
44Skill Claude Code
Semantic search across codebase using LEANN vector index.
not rated 3.9k
+2 8mo ago A SkillSpector: pass 14 tokens
original MIT
Skill Claude Code
Route through domain routers before using individual tools. Routers abstract tool selection.
not rated 3.9k
+2 8mo ago A Socket: passSnyk: passSkillSpector: pass 9 tokens
original MIT
Skill Claude Code
Use the most token-efficient search tool for each query type.
not rated 3.9k
+2 8mo ago A Socket: passSnyk: passSkillSpector: pass 8 tokens
original MIT
search-tools
47Skill Claude Code
When searching code, use this decision tree.
not rated 3.9k
+2 8mo ago A Socket: passSnyk: passSkillSpector: pass 7 tokens
original MIT
Skill Claude Code
AST-based code search and refactoring via ast-grep MCP.
not rated 3.9k
+2 8mo ago A Socket: passSnyk: passSkillSpector: pass 17 tokens
original MIT
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: