benchmark
Run SWE-bench Lite benchmarks against agtx coding agent workflows. Guides setup, configuration, execution, evaluation, and reporting.
Self-review completed work. Check for correctness, edge cases, and code quality. Write review to .agtx/review.md and stop.
$ npx -y skills add fynnfluegge/agtx --skill agtx-review --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/agtx-reviewContext preview
The summary Claude sees to decide when to auto-load this skill.
Self-review completed work. Check for correctness, edge cases, and code quality. Write review to .agtx/review.md and stop.
name: agtx-review description: Self-review completed work. Check for correctness, edge cases, and code quality. Write review to .agtx/review.md and stop.
You are in the **review phase** of an agtx-managed task.
1. Pick the scope. If `.agtx/reviewed-at` exists, a prior review covered everything up to that commit: review `git diff $(cat .agtx/reviewed-at)..HEAD` plus `git status --short` for uncommitted work, and read the prior `.agtx/review.md` — check its points were addressed, don't re-review the branch. Otherwise review all changes made during execution: `git diff HEAD` (staged+unstaged) and `git log --oneline $(git merge-base HEAD origin/HEAD)..HEAD` for your commits. Do NOT diff against `main` or `origin/main` — those may include unrelated upstream history. Either way `??` lines in `git status --short` are untracked files that `git diff` never shows — read them in full. 2. Check for:
3. Fix any issues you find 4. Commit them. `git add -A`, commit, verify `git status --short` clean. Uncommitted = destroyed at task completion; merge is blocked until clean.
Write your review to `.agtx/review.md` in the **current working directory** with these sections:
Findings from your review — what looks good, what was fixed, any concerns.
Either `READY` (good to merge) or `NEEDS_WORK` (with explanation of remaining issues).
After writing `.agtx/review.md` (in the current working directory):
Terse. No pleasantries. Fragments OK. Short synonyms. Code exact. Status updates: one line. Pattern: [what] [why]. Done.
The terminal-native agentic development environment for 10x productivity. A blackboard for coding agents - One shared board. A fleet of agents.
Run SWE-bench Lite benchmarks against agtx coding agent workflows. Guides setup, configuration, execution, evaluation, and reporting.
Execute an approved implementation plan. Implement the changes, then write a summary to .agtx/execute.md and stop.
Plan a task implementation. Analyze the codebase, create a detailed plan, write it to .agtx/plan.md, then stop and wait for user approval before making any…
Explore the codebase to understand a task before planning. Write findings to .agtx/research.md and stop. This is a read-only exploration — do not modify any…
Enter brainstorm mode to explore a feature or enhancement idea. Stays in discussion mode only — no planning, no implementation. Use /agtx:sweep when ready to…
One-shot a whole project on an agtx board: decompose the goal, run every task unattended, unblock the workers, and merge each one. Use when the user wants a…