AGENTS.md (Skills)

OVERVIEW

Control plane for Obsidian vault operations. Manages skill-based routing to specialized reference files. Reduces context overhead by loading only relevant guidance.

STRUCTURE

skills/
├── AGENTS.md (this file)
└── obsidian/
    ├── AGENTS.md (skill-level logic)
    ├── SKILL.md (main router)
    └── references/
        ├── AGENTS.md (reference-level guidance)
        ├── integrity-sweep.md
        ├── para-routing.md
        ├── people-style-profile.md
        ├── read-workflow.md
        ├── task-conventions.md
        ├── transcript-ingestion.md
        └── write-routing.md

SKILL ROUTING

• Entrypoint: agents/obsidian.md delegates to skills/obsidian/SKILL.md.
• Router: SKILL.md analyzes the task and selects the appropriate reference.
• Delegation:
  • Reading/Querying -> read-workflow.md
  • Writing/Editing -> write-routing.md
  • Organization -> para-routing.md
  • Post-Edit Checks -> integrity-sweep.md
  • Transcripts -> transcript-ingestion.md
  • Tasks -> task-conventions.md
  • People Notes -> people-style-profile.md

CONVENTIONS

• Thin Router: SKILL.md should contain minimal logic; delegate to references.
• Context Efficiency: Only load the specific reference file needed for the current step.
• Separation of Concerns: Keep markdown formatting in obsidian-markdown skill; keep vault logic here.
• Hierarchical Guidance:
  • Root AGENTS.md: Project-wide overview.
  • skills/AGENTS.md: Skill-level coordination (this file).
  • skills/obsidian/AGENTS.md: Obsidian-specific skill logic.
  • skills/obsidian/references/AGENTS.md: Reference-level guidance.

ANTI-PATTERNS

• Logic Duplication: Don't repeat rules across files.
• Confident Guessing: Ask if identity or routing is ambiguous.
• Silent Deletion: Never delete user content without request.
• Unplanned Reading: Don't read notes without a plan.
• Monolithic Prompts: Avoid putting all logic in the agent file.

REFERENCES

• Root AGENTS.md
• Obsidian Skill AGENTS.md