benchling-integration Eval
==========================

Status: failed
Score: 72/100
Risk: high
Decision: do_not_auto_install
Policy: block
Reason: Permission surface: secrets or environment access, shell or command execution

Install:
npx skills add K-Dense-AI/scientific-agent-skills --skill benchling-integration

Required checks:
- PASS Task fit: Task wording matches this skill metadata.
- PASS Install path: Install handoff is available.
- PASS Install command safety: standard package or runtime install path
- WARN Trust score: Potentially useful, but at least one trust signal needs human inspection.
- WARN Audit score: Needs review
- WARN Agent safety gate: Sparse or mixed signals. Useful for discovery, but not for autonomous installation.
- PASS License clarity: MIT
- FAIL Permission surface: secrets or environment access, shell or command execution

Warnings:
- Trust score: Potentially useful, but at least one trust signal needs human inspection.
- Audit score: Needs review
- Agent safety gate: Sparse or mixed signals. Useful for discovery, but not for autonomous installation.
- High-risk permission hints: Shell or command execution, Secrets or environment access
- Dependency or permission surface needs review
- Permission surface may require sandboxing
- SKILL.md lacks an explicit limitations or troubleshooting section, so safe operating boundaries are not fully described.
- The metadata marks BENCHLING_API_KEY as primaryEnv while also listing it as not required, which could confuse agents about which auth path to use.
- The skill does not warn agents to treat Benchling API responses as untrusted data, which is relevant for prompt-injection safety when processing external records.
- Permission surface needs review: secrets or environment access, shell or command execution
- Dependency/runtime risk: command execution surface, credential or environment access
- Permission surface: secrets or environment access, shell or command execution

Validation plan:
1. Inspect repository, README/SKILL.md, license, and recent commits before production use.
2. Install in an isolated workspace or sandbox with no production secrets available.
3. Run the smallest representative task and record files touched, commands run, network access, and outputs.
4. Compare the selected skill against at least one alternative when the eval status is review or failed.
5. Promote only after the agent reports a successful verification result and unresolved warnings are accepted.

Do not use when:
- teams that need a vendor-supported SLA
- production agents without a repository review
- SKILL.md lacks an explicit limitations or troubleshooting section, so safe operating boundaries are not fully described.
- High-risk permission hints: Shell or command execution, Secrets or environment access
- Dependency or permission surface needs review
- Permission surface may require sandboxing
- The metadata marks BENCHLING_API_KEY as primaryEnv while also listing it as not required, which could confuse agents about which auth path to use.
- The skill does not warn agents to treat Benchling API responses as untrusted data, which is relevant for prompt-injection safety when processing external records.

URLs:
- Skill: https://www.openagentskill.com/skills/k-dense-ai-benchling-integration
- Audit: https://www.openagentskill.com/skills/k-dense-ai-benchling-integration/audit
- JSON: https://www.openagentskill.com/api/agent/evals?slug=k-dense-ai-benchling-integration