Community

vendor-evaluation

Open Agent Skill · documents · works with Claude Code, Codex, Cursor, and compatible agents

Run structured vendor evaluations and scorecards for build-vs-buy decisions. Use whenever the user compares vendors, shortlists tools, runs a bake-off, or needs a fair scorecard across products for a purchase decision.

View on GitHub

Vendor Evaluation

Score what matters: must-haves, weighted wants, risks, and total cost — not demo polish alone.

Workflow

  1. Decision, constraints, must-haves vs nice-to-haves.
  2. Criteria + weights (security, cost, UX, ops, lock-in).
  3. Candidate list and disqualify early misses.
  4. Evidence: docs, trials, references — not brochure claims.
  5. Scorecard + recommendation + risks.
  6. Pilot plan if close.

Output format

## Evaluation: <category>
### Must-haves
### Scorecard
| Criterion | Weight | Vendor A | Vendor B | Notes |
### TCO notes
### Risks / lock-in
### Recommendation
### Pilot plan

Rules

  1. Never invent SOC2/pricing; mark unknown.
  2. Weight security for sensitive data.
  3. Include exit cost / data portability.
  4. Separate build option when relevant.
  5. Document who scored and when.

Edge cases

  • Incumbent bias: re-score fairly.
  • Open source + support vendor: split product vs support.
  • Multi-year contracts: renegotiation leverage notes.