superpowers-sync-upstream
Skill101 starsChanged 30 days agoMITView on GitHub
Sync this port with upstream obra/superpowers or Reasonix main-v2: detect drift, re-port, bench-gate
In this repository12 other files it gives its agents.
- AGENTS.mdAGENTS.md
- superpowers-brainstormingSkill
- superpowers-diagnosing-superpowersSkill
- superpowers-executing-plansSkill
- superpowers-finishing-a-development-branchSkill
- superpowers-receiving-code-reviewSkill
- superpowers-systematic-debuggingSkill
- superpowers-test-driven-developmentSkill
- superpowers-using-git-worktreesSkill
- superpowers-verification-before-completionSkill
- superpowers-writing-plansSkill
- superpowers-writing-skillsSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 0 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 12
Read UPSTREAM.json for the pinned commits and watched paths, then see what moved:
- Very specific: line 14
Do not run .github/scripts/check-upstream-drift.mjs for a read-only check: it can create or update the tracking issue.
- Very specific: line 44
Update the affected skills, AGENTS.md, and bench config.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Sync Upstream
Re-aligns this port with its two upstreams and re-pins UPSTREAM.json. Run when asked to sync or update from upstream, or after the upstream-drift CI issue fires.
1. Detect
Read UPSTREAM.json for the pinned commits and watched paths, then see what moved:
Do not run .github/scripts/check-upstream-drift.mjs for a read-only check: it can create or update the tracking issue. Inspect the pinned SHAs with git ls-remote or gh api repos/<owner>/<repo>/commits/<branch>, then compare the two endpoints with Git. Use the issue-writing script only when updating the tracker is intended.
For full diffs, shallow-fetch and compare from the pins:
Fetch both the pinned commit and current default-branch head into a temporary blobless clone, then run git diff --no-renames <pin> <head> -- skills/. A GitHub compare API response with 300 files may be truncated; do not treat it as complete.
For Reasonix, diff the pin against main-v2 over internal/tool internal/hook internal/skill internal/frontmatter internal/command. Inspect actual tool registration and implementation together; a declared identity alone does not prove a callable tool exists.
Network blocked? web_fetch the GitHub compare page: <repo>/compare/<pinned>...<branch>, and label incomplete file lists. This step is read-only; dispatch it to a task subagent when appropriate and keep only the drift report to save context.
2. Classify content drift
For each changed obra skill, decide edited, new, or removed.
- New skill: port it only if no native Reasonix tool already covers it (
task,review,wait,explore). Otherwise add it toUPSTREAM.jsonretiredwith a one-line reason. - Removed upstream skill: consider retiring the port equivalent.
3. Re-port
Load superpowers-writing-skills. For each edited or to-be-ported skill, apply the Adaptations rules: snake_case tool names (read_file, edit_file, bash), description rewritten as a forceful imperative within the 130-char index budget, references/* auto-fold, native-tool substitution, path renames (docs/reasonix/...). Preserve the disciplines verbatim: Iron Laws, red-flag tables, RED-GREEN-REFACTOR. Keep bodies lean enough for the flash model.
4. Classify harness drift
Map each changed contract file to the port surface it touches:
internal/tool: tool names in every skill body andbench/reasonix.tomlinternal/hook:AGENTS.mdnotes and any hook docsinternal/frontmatter: all skill frontmatterinternal/skill,internal/command: layout, loading, and/nameassumptions
Update the affected skills, AGENTS.md, and bench config. If the change invalidated the local reasonix binary, rebuild it before trusting bench.
5. Bench-gate
node --test bench/structural.test.mjs(structure of every SKILL.md)node bench/bench.mjs(invocation; prior cases must still pass)- exec-fidelity for any re-ported discipline skill, vs
bench/BASELINE.json
Any structural failure or exec score below baseline blocks acceptance. Fix the skill body; do not lower the baseline.
6. Finalize
Bump UPSTREAM.json commits and dates to the synced HEADs, write a short report (changed, ported, retired, bench result), and commit. Close the open upstream-drift issue only after the synchronized changes reach main (for example, via the PR's closing keyword).
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

