skills.rest
wimpysworld
wimpysworldOFFICIAL
Explore all skills by this author →
WebsiteGitHub

review-report-path

Determines durable per-user storage paths for review and audit reports.

716|31|Updated Mar 13, 2023
Search Tags:#AI Agent#file path#report storage#Git worktree#Code review#Pull request#xdg state
One-click install
npx skills add https://github.com/wimpysworld/nix-config --skill review-report-path
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: review-report-path
Source: https://github.com/wimpysworld/nix-config/tree/main/home-manager/_mixins/agentic/assistants/skills/review-report-path
Command: npx skills add https://github.com/wimpysworld/nix-config --skill review-report-path

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Review, audit, and analysis commands need a consistent, durable location to write reports and a reliable way to find reports written earlier, without losing files to ephemeral /tmp storage or colliding with parallel reviews.

Core Features & Use Cases

  • Durable Report Storage: Writes reports to per-user state storage under ${XDG_STATE_HOME:-~/.local/state}/agent-reviews instead of /tmp, so reports survive process exits.
  • Target Slug Rules: Derives collision-free slugs for pull requests, issues, commits, branches, worktrees, and local files so concurrent reviews never overwrite each other.
  • Run Isolation and Lookup: Creates a unique run directory per invocation via mktemp and provides shell snippets to locate past reports across all runs.
  • Use Case: An AI agent reviewing pull request #123 writes its report to agent-reviews/nix-config/pr-123/run-20240101T000000Z-XXXXXX/review.md, and a later session can retrieve that exact report by searching the target directory.

Quick Start

Ask the agent to write the review report for the current pull request using the standard report path rules.

Frequently Asked Questions about review-report-path

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
Where should AI agents write code review reports?▼

Review reports should go to durable per-user state storage at ${XDG_STATE_HOME:-~/.local/state}/agent-reviews/<project>/<target>/<run-id>/. This avoids /tmp, which can be private to a fenced process and disappear after it exits.

How do I keep parallel code reviews from overwriting each other?▼

Derive a target slug from what is under review, such as pr-<number> for pull requests or commit-<short-sha> for commits, then create a unique run directory per invocation with mktemp -d. Two concurrent reviews then never write to the same file.

How do I find a review report written in an earlier session?▼

Derive the same project and target directory in per-user state storage, then iterate over all run-* directories looking for the report file. When several runs match, ask the caller to select an exact run rather than picking the newest.

Why not write agent review reports to /tmp?▼

A fenced or sandboxed process can have private temporary storage that disappears when the process exits, losing the report. Per-user XDG state storage persists across sessions and keeps reports outside the repository.

Can review reports be committed to the git repository?▼

No. Reports must stay outside the repository in per-user state storage and must never be committed. Run directories and reports are also never deleted or reused.

Related Skills

Explore More Software Engineering Skills→
nsheapsnsheaps
3

git-worktree

Generates Git commands for safe branch matching in worktrees.

Community
Intermediate
oduffy-delphioduffy-delphi
5

handoff-archival

Move old handoffs to archive/ after validating .gitignore safety and git branch activity.

Community
Advanced
liminal-ailiminal-ai

review-synthesis

Consolidate multi-pass code review findings into a verified single report.

Official
Advanced
stacklokstacklok
2.0k

pr-inline-review

Submit inline review comments and code suggestions to GitHub Pull Requests via gh CLI.

Official
Intermediate