Agent Claude Code
Read-only search agent for broad fan-out searches — when answering means sweeping many files, directories, or naming conventions and you only need the conclusion, not the file dumps. It reads excerpts rather than whole files, so it locates code; it doesn't review or audit it. Specify search breadth — 'medium' for…
Agent Claude Code
Implementation requiring judgment — feature work, bug fixes, refactors with design decisions, integration work. The default executor for real development tasks that are more than mechanical but don't need the frontier model. Give it the goal, constraints, and done-criteria; it makes reasonable local design decisions…
Agent Claude Code
Mechanical execution of fully-specified work — pattern-based refactors and renames, writing tests that follow existing conventions, documentation updates, bulk multi-file edits from an explicit spec, running test suites and fixing trivial failures. Use when the task needs no design decisions; give it a complete spec…
Agent Claude Code
Explore the codebase, find existing patterns, and gather context before implementation. Use when searching for prior art or answering architectural questions.
Agent Claude Code
Read-only reconnaissance. Use for any search, lookup, or 'where/how is X' question that requires no judgment — locating files, symbols, usages, config values, or summarizing how something works across the codebase. Returns concise findings with file:line references. Cheapest way to gather facts; prefer it over reading…
Agent Claude Code
Isolated SDLC pipeline lane worker. Spawned by the sdlc-dispatch scheduled task to execute one prompts/sdlc/ .md pass. Deliberately has NO agent-spawning tool: all owner work is done inline.
Agent Claude Code
Security-sensitive implementation and analysis — authentication/authorization, secrets handling, crypto usage, input validation, hardening, dependency vulnerability triage, security-relevant code review. Use for ANY task where the word 'security' applies, instead of executor or the main session.
Agent Claude Code
Fresh-context adversarial verification of completed work. Use after any non-trivial change, before reporting it done — give it the claimed outcome and the diff/paths, and it independently tries to refute the claim by exercising the code, running tests, and probing edge cases. Returns CONFIRMED or REFUTED with…
Hook Claude Code
Runs before the agent uses a tool for Bash, Read, Glob, Grep and Bash tool calls, executing graphify-nudge.py and pii-guard.py via python3 (2 commands). From meridun/pemr.
Hook Claude Code
Runs when you submit a prompt, before the agent sees it, running an inline shell check. From meridun/pemr.
Settings file Claude Code
Agent settings declaring 2 hook events (PreToolUse, UserPromptSubmit).
Skill Claude CodeCodex
Create and maintain Agent Skills for this repository. Use when adding new skills, updating existing skills, or registering skills in copilot-instructions.md.
Skill Claude CodeCodex
Harvest session learnings into the documentation tier system and audit/reorganize the tiers (L1 copilot-instructions, L2 skills, L3 docs, agents, memory). Use when a session ends with discoverable information ("harvest this session", "update the doc tiers", "capture what we learned"), or when auditing, rebalancing, or…
Instructions file GitHub Copilot
Copilot instructions for meridun/pemr, covering copilot instructions, core principles, memory vs documentation, documentation tiers and compound tasks (load skills sequentially, not all at once).
Instructions file CodexOpenCode
AGENTS.md instructions for meridun/pemr, covering agents.md — the pemr agent contract, client configuration, tool surface, must rules and 1. extraction naming.
Instructions file
Claude Code instructions for meridun/pemr, covering claude code notes, skills and agents, graphify and token wrappers.
MCP server Claude CodeCodexCursor +2
MCP server "pemr" as configured in meridun/pemr. Runs locally from the pemr Python package.