evaluate-tool

name	evaluate-tool
description	Evaluate a tool or integration for the productivity system. Researches the tool, compares alternatives, assesses fit, and creates an evaluation note in Notes/. Use when the user says 'evaluate', 'should we use', 'compare tools', 'assess this integration', or 'what about adding'.
argument-hint	[tool or integration to evaluate]
context	fork
agent	system-architect

Evaluate: $ARGUMENTS

Current integrations

!cat .claude/rules/system-architecture.md 2>/dev/null || echo "(Could not read system-architecture rule)"

Previous evaluations

!obsidian search query="tool-evaluation" vault="{{VAULT_NAME}}" limit=5 2>/dev/null || echo "(Could not search vault — Obsidian may not be running. Use Grep/Glob to search files directly.)"

Your task

Evaluate $ARGUMENTS for potential integration into the productivity system.

Follow this workflow:

1. Understand the candidate — What is it, what problem does it solve, how does it connect (MCP server, CLI, API, plugin)? Does it overlap with existing tools?

2. Research — Use WebSearch for:

   • Official documentation and capabilities
   • Adoption signals (community size, maintenance activity)
   • Claude Code / MCP compatibility
   • Known limitations and gotchas Use WebFetch to deep-read the 2-3 most relevant sources.

3. Check vault knowledge — Search the vault for:

   • Existing notes about this tool or similar tools — use mcp__qmd__query with lex (tool name) + vec (problem it solves) and intent describing the evaluation context. See the qmd rule for query construction patterns.
   • Previous evaluations in Notes/
   • Check your memory for past evaluation context

4. Compare alternatives — Assess against current tools:

   • Feature coverage overlap and gaps
   • Integration complexity (how much system change is needed?)
   • Maintenance burden (who updates it? how often?)
   • UX impact (does it simplify or complicate the user's workflow?)

5. Assess system impact — If adopted, what changes?

   • Which rules, skills, or hooks need updates?
   • New protected paths needed?
   • New agent or skill required?
   • Risk if the tool is abandoned or breaks?

6. Create evaluation note — In Notes/ with:

   ---
   categories: ["[[Research]]"]
   status: active
   areas: []
   tags: [tool-evaluation]
   created: YYYY-MM-DD
   ---
   

   Body: overview, comparison matrix, integration plan, recommendation, revisit triggers.

7. Link and commit — Verify categories and areas properties. Commit the evaluation note.

Output

Return:

• Recommendation — Adopt / Defer / Reject (with clear rationale)
• Key findings — 3-5 bullets of the most important discoveries
• Evaluation note — Path to the created vault note
• Integration effort — Low / Medium / High
• Revisit triggers — Conditions that should prompt re-evaluation