ds-unstick — Claude Code skill: what it does, how to install
ds-unstick — a Claude Code skill breaks repeating failure loops by testing core assumptions. It runs when the same error returns after two failed fix attempts.
MIT 566 installs ★ 320 updated 2026-10-28
npx skills add aicodeking/deepseek-stack --skill ds-unstick -a claude-codeWhat ds-unstick does
The skill tackles stubborn bugs where minor tweaks keep failing. It stops further edits, lists failed attempts with exact errors, and isolates the underlying assumption shared by all failures.
A direct probe checks whether that assumption holds, and if false, points straight to the real issue. When the assumption holds, the skill lists three alternative approaches from different layers or libraries, picks one, and cleans up old edits with git stash or git checkout before applying the new plan.
Progress and tested assumptions are logged in PROGRESS.md so later sessions avoid repeating dead ends.
What it helps with
- break out of repetitive fix loops
- test core assumptions behind failing code
- clean up failed edits using git
- record verified approaches in PROGRESS.md
Example prompts for the agent
- /ds-unstick
- Run /ds-unstick, we are looping on this bug
- Use /ds-unstick to break out of this failure
How to install ds-unstick
1Skills CLI (Vercel)
npx skills add aicodeking/deepseek-stack --skill ds-unstick -a claude-code2Manually: put the files in place
~/.claude/skills/ds-unstick/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/ds-unstick/
1Skills CLI (Vercel)
npx skills add aicodeking/deepseek-stack --skill ds-unstick -a codex2Manually: put the files in place
.agents/skills/ds-unstick/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add aicodeking/deepseek-stack --skill ds-unstick -a cursor2Manually: put the files in place
.cursor/skills/ds-unstick/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install aicodeking/deepseek-stack skills/ds-unstick2Manually: put the files in place
.github/skills/ds-unstick/SKILL.md1Skills CLI (Vercel)
npx skills add aicodeking/deepseek-stack --skill ds-unstick -a gemini-cli2Platform CLI
gemini skills install https://github.com/aicodeking/deepseek-stack --path skills/ds-unstickManually: put the files in place
.devin/skills/ds-unstick/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add aicodeking/deepseek-stack --skill ds-unstick -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about ds-unstick
How to install ds-unstick in Claude Code?
Run npx skills add aicodeking/deepseek-stack --skill ds-unstick -a claude-code and reference ~/.claude/skills/ds-unstick/SKILL.md.
What does ds-unstick do?
It stops repeating failed fixes, tests the underlying assumption directly with a probe, and proposes three alternative approaches from different layers.
What is needed to run ds-unstick?
You need Node.js installed to run npx commands and Git for repository cleanup.
Source text
SKILL.md · MIT
--- name: ds-unstick description: "Break out of a repeating failure: name the assumption behind all the failed fixes, test it directly, list three other approaches, pick one, and restart clean if needed. Use when the same error or wrong behavior comes back after two fix attempts, or when you notice you are retrying small variations. Invoke as /ds-unstick." --- # ds-unstick When two fixes fail the same way, the problem is usually a wrong belief, not a wrong edit. Find that belief and test it. ## Steps 1. Stop editing. Do not try a third variation of the same fix. 2. Write down each failed attempt in one line: what you changed and what happened. Include the exact error text. 3. Write the assumption that all of those attempts shared. Typical examples: "this file is the one that runs", "the config is loaded", "the function receives a string", "the cache is cleared", "the error comes from my code". Name one, in one sentence. 4. Test that assumption directly with the cheapest probe that gives a yes or no answer: a print statement, a log line, a `which` or `--version`, a grep for the real call site, a minimal script, or a read of the docs for the exact version installed. Paste the result. 5. If the assumption is false, you have found the real problem. Plan the fix from the new fact. 6. If the assumption holds, list 3 different approaches. "Different" means another mechanism, library, layer or data path, not a tweak of the same edit. Give each approach one line on cost and risk. 7. Pick one approach and say why in one sentence. 8. Restart from a clean state if the failed attempts left debris: revert them with git (`git stash` or `git checkout -- <files>`), then apply only the chosen approach. 9. Record the false assumption or the chosen approach in PROGRESS.md (see ds-context-reset) so a later session does not repeat the attempts. 10. Return to ds-verify-loop with the new approach. ## Stop when - The probe in step 4 explains the failure, or the new approach passes its check. - Two rounds of this skill have failed. Stop and report to the user: the attempts, the assumptions you tested, and the evidence. Ask for direction. ## Rules - Do not repeat a fix that has already failed, even with small changes. - Evidence beats reasoning. If a probe takes a minute, run it. - Keep the failed-attempt list, because it is the most useful part of the report. - Use a clean state before a new approach, so that old edits cannot hide the result.
© aicodeking · MIT · aicodeking/deepseek-stack