PugzUI

PugzUI/axom

Instructions file CodexOpenCode

AGENTS.md instructions for PugzUI/axom: See docs/agents/INDEX.md for the full agent reference, tool index, and setup instructions.

PugzUI/axom

Agent

Agent "INDEX" from PugzUI/axom, covering axom agent documentation index, standard workflow, quick reference cards, discover (find) and exec (run).

PugzUI/axom

Agent

Common issues and solutions for Axom MCP.

PugzUI/axom

Skill Claude CodeCodex

Deep analysis. Actionable output. Memory integration. Analysis without storage is wasted. Chain findings to memory. Use reflex for heuristics.

PugzUI/axom

Skill Claude CodeCodex

Map the unknown before you act. Violently effective recon. No blind discovery. Filter aggressively. Chain to memory.

PugzUI/axom

Skill Claude CodeCodex

When the problem is complex and context is thin: Dispatch a subagent to gather context in parallel while you solve.

PugzUI/axom

Skill Claude CodeCodex

Atomic chains. Tool abstraction. Zero wasted steps. Read → transform → write in one call. 3–7 steps ideal.

PugzUI/axom

Skill Claude CodeCodex

Memory tool that challenges itself, other agents, and users. Re-think. Seek optimal solutions. Improves creativity and focus. Use for self-reflection, knowledge exploration, deepthink, dreams.

PugzUI/axom

Skill Claude CodeCodex

When search returns reflex memories: Implement from experience, verify, and update the reflex if verification fails.

PugzUI/axom

Skill Claude CodeCodex

When to use: At session start or when you need to refresh knowledge of available tools and skill guides.

PugzUI/axom

Skill Claude CodeCodex

Format mastery. Pipeline glue. Data flow. Connective tissue of chains. Read → Transform → Write.

PugzUI/axom

MCP server Claude CodeCodexCursor

MCP server "axom-mcp" as configured in PugzUI/axom. Runs locally from the axom-mcp Python package.