fresh-eyes-review

Solid

Fresh-eyes review of a changeset by a fresh-context agent — catches regressions and correctness issues the authoring context reads past.

Code & Development 26 stars 3 forks Updated today MIT

Install

View on GitHub

Quality Score: 87/100

Stars 20%
48
Recency 20%
100
Frontmatter 20%
70
Documentation 15%
100
Issue Health 10%
50
License 10%
100
Description 5%
100

Skill Content

# Fresh-eyes review A context that produced a change reads its intent, not its text, so same-context review misses what a fresh reader would catch. The fix is procedural: a reviewer whose context holds only the artifacts. ## Workflow 1. **Resolve the inputs.** The changeset: whatever the invocation names — a branch, a commit, a diff range, a draft vs its original. Given none, infer it from the session — usually the work just finished, committed or not; no VCS required. With no session context to draw on, fall back to the current git diff; when that too yields nothing, ask the user what to review. Pin the changeset as concretely as the environment allows — a diff or commit range where one exists, otherwise the touched files, with their prior state when reconstructable. Alongside it, a short statement of what the change is supposed to achieve, when one exists (the task as stated, a PR or ticket description); when this session authored the change, never include the session's own reasoning, plan, or messages — leaked rationale recreates the blindness the fresh context exists to remove. Done when changeset and intent are pinned down and free of authoring context. 2. **Confirm the prompt.** Assemble the reviewer prompt — changeset, intent, and the mandate and exclusions below — show it to the user verbatim, and wait for approval; fold any doubt about an inferred changeset into the proposal rather than asking separately. Text emitted before a tool...

Details

Author
eai-org
Repository
eai-org/agent-toolkit
Created
2 months ago
Last Updated
today
Language
Shell
License
MIT

Bundled in these plugins

Similar Skills

Semantically similar based on skill content — not just same category