greploop Eval report.

Validation plan

What the agent should do next

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

Do not use when

Conditions that require another skill

• teams that need a vendor-supported SLA
• production agents without a repository review
• The SKILL.md does not explicitly describe how to fetch the list of unresolved review comments or how to extract the Greptile confidence score from the review output. This is a critical step in the loop (fixing actionable comments) but is left implicit.
• High-risk permission hints: Shell or command execution
• Permission surface may require sandboxing
• Inputs (PR/MR/CL number, max iterations) are used in shell commands without explicit quoting or validation guidance. An agent following the instructions literally could be vulnerable to command injection if the inputs are maliciously crafted.