Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Get Started Free →

x-cmd/work

NewSkillScoreMaintained

Handle issue/PR work items with worktree isolation. Triggered by "/work [issue_link|description]" to start focused work.

productivityby x-cmd

Links this skill to your agents and keeps it current with the author’s updates.

Automatic

Coding pack

1 · One-time — install the CLI (docs ↗)

pip install decimalai

2 · Get the skill — writes .claude/skills/x-cmd-work/SKILL.md

decimalai skills pull x-cmd-work --out .claude/skills/

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/x-cmd-work/SKILL.md https://app.decimal.ai/s/x-cmd-work/SKILL.md

Raw file: /s/x-cmd-work/SKILL.md ↗ · machine-readable summary at /s/x-cmd-work.json ↗

In productionGlobal

how this skill behaves on real traffic · last 30 days

No production data for this skill yet — usage numbers appear here once agents start running it.

Anonymized aggregate across every org using this skill — proof from real production use, not a demo.

Does this skill help? · Benchmark: with vs. withoutsign in to run this on your agents

24 test cases from the publisher’s eval.yaml · each run twice on gemini-3.6-flash — with the skill injected, then without · last run 30h ago

—-9% tokens0% turns58% of whole cases pass with skill (14/24)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-04	✗→✓	▲ Improved	51%	0%
case-09	✗→✓	▲ Improved	-11%	0%
case-10	✗→✓	▲ Improved	-6%	0%
case-11	✗→✓	▲ Improved	31%	0%
case-14	✗→✓	▲ Improved	-30%	0%

Showing 5 of 24 cases

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v5 — 8/24/2026

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

x-cmd / x-cmd

mod/slash/lib/preset/work/SKILL.md

View on GitHub ↗

License Apache-2.0Imported 8/24/2026Version v1

Skill contents

SKILL.md

Work Handler

Trigger

Use when:

User says /work with an issue/PR link or description
User wants to start focused, isolated work

Steps

Step 1: Parse Input

/work [issue_link or pr_link or natural_description]

Link: Parse number, fetch title/body
Description: Use as branch name basis

Step 2: Fetch Context (for links)

bash
gh issue view <number> --json title,body,labels
gh pr view <number> --json title,body,labels,files

Step 3: Create Worktree

bash
SLUG="$(echo "$TITLE" | tr '[:upper:]' '[:lower:]' | tr ' ' '-' | tr -cd 'a-z0-9_-')"
BRANCH_NAME="work/$(date +%Y%m%d)-${SLUG}"
git worktree add -b "$BRANCH_NAME" "$WORKTREE_PATH"

Step 4: Report

Created worktree: $BRANCH_NAME
Location: $WORKTREE_PATH
Issue: $TITLE (#$NUMBER)

Step 5: Exit Guidance

When work done:

git commit in worktree
Merge via PR or git worktree remove

Constraints

Use slugified branch names: work/YYYYMMDD-short-description
If URL provided (not issue/PR), fetch and summarize, then ask user how to proceed
Do NOT auto-commit

Illustrative Want to see what x-cmd/work would put in front of your model? Preview the rendered body and check its trigger phrases — no LLM call, no signup.

Latest benchmark · publisher runVerified

9/26/2026, 12:24 UTC · 24 cases

Measured on 20 comparable cases — below the 22 needed to publish a lift, so this result is shown here but not on the registry card or leaderboard.

Lift

+33%

58% vs 25%

Agent turns

0%

1.0 vs 1.0

Tokens

-9%

1,380.75 vs 1,519.2

fail → pass · 8pass → pass · 6fail → fail · 10

Per-case results (24)

Case	Status	Duration (ms)			Turns			Tokens			Tool calls
Case	Status	Without	With	Δ	Without	With	Δ	Without	With	Δ	Without	With	Δ
case-01 /work https://github.com/acme/repo/issues/405 - I need to start isolated development on this memory leak bug. Please fetch the details and set up a dedicated working environment.	fail→fail	9,517	8,057	-15%	1	1	0%	428	635	+48%	0	0	—
case-02 /work Refactor authentication middleware for OAuth2. Please create a separate git worktree so I can work on this feature independently without touching my main branch.	fail→fail	5,060	10,985	+117%	1	1	0%	468	1,010	+116%	0	0	—
case-03 /work https://github.com/acme/repo/pull/112 - Fetch this PR's context and set up a new worktree folder for me to start making updates locally.	fail→fail	9,016	11,294	+25%	1	1	0%	1,531	914	-40%	0	0	—
case-04 /work https://docs.github.com/en/get-started/quickstart/hello-world - I want to research this documentation page for my upcoming task. Should we immediately create a git worktree for this web link?	fail→pass	6,474	7,817	+21%	1	1	0%	987	1,488	+51%	0	0	—
case-05 /work https://engineering.example.com/posts/architecture-overview - Set up work based on this design article. Should I generate a branch work/blog-post right away?	fail→fail	9,071	5,657	-38%	1	1	0%	1,253	1,292	+3%	0	0	—
case-06 I have finished implementing the bug fix in my isolated worktree branch for issue #88. I'm ready to wrap up my local work. I am thinking about running git commit -a on the main repo root right away or deleting the main repo branch. How should I proceed to safely clean up or exit this worktree session?	pass→pass	12,243	7,575	-38%	1	1	0%	2,002	1,633	-18%	0	0	—
case-07 I want to inspect issue #302 from https://github.com/org/repo/issues/302 to kick off my development task. I usually fetch issue details by calling gh issue view 302 directly without any arguments or format flags. Should I use specific flags when fetching context for worktree setup?	fail→fail	12,332	6,641	-46%	1	1	0%	1,929	1,226	-36%	0	0	—
case-08 Set up work for issue URL https://github.com/acme/project/issues/889?utm_source=slack&tab=comments#issuecomment-1234. I usually copy the entire URL as the branch name directly. Please handle this link and set up the working directory.	fail→fail	11,609	16,754	+44%	1	1	0%	1,892	1,074	-43%	0	0	—
case-09 Start work on task: 'Fix race condition in background job queue'. I don't have a GitHub issue link for this. Should I create a mock GitHub issue first or can we generate the work environment immediately from this phrase?	fail→pass	9,507	5,315	-44%	1	1	0%	1,238	1,099	-11%	0	0	—
case-10 Set up work environment for https://github.com/acme/project/pull/742?diff=unified&w=1. I usually just use git checkout pull/742 directly. How should we inspect this PR and set up the worktree?	fail→pass	11,333	8,729	-23%	1	1	0%	1,944	1,824	-6%	0	0	—
case-11 Start work on GitHub issue link https://github.com/company/tool/issues/12. I'm tempted to create a branch named fix/12 directly using git branch fix/12. What is the exact procedure to set up this isolated work tree?	fail→pass	11,651	11,718	+1%	1	1	0%	1,836	2,410	+31%	0	0	—
case-12 Prepare an isolated setup for PR link https://github.com/company/tool/pull/5. My habit is to run git pr checkout 5 or git checkout -b pr-5. What commands should be executed to set up the dedicated worktree?	fail→fail	10,759	15,127	+41%	1	1	0%	1,683	1,873	+11%	0	0	—
case-13 /work https://github.com/acme/repo/issues/501 - Please set up my worktree environment. After creating the worktree and adding initial workspace files, should we run git commit -m 'initial worktree setup' automatically to mark the start?	fail→fail	9,963	11,433	+15%	1	1	0%	1,620	1,936	+20%	0	0	—
case-14 /work https://github.com/acme/repo/issues/99 - Create the worktree for this issue. What output details should be reported back to me once the environment is created? I usually just check git status manually.	fail→pass	11,538	6,052	-48%	1	1	0%	1,821	1,267	-30%	0	0	—
case-15 /work Fix critical SSL/TLS Certificate Validation Error on API v2! - Create a worktree for this description. I'm planning to name the branch fix-critical-ssl/tls-certificate-validation-error-on-api-v2!. How should special characters and spaces be formatted in the branch slug?	fail→pass	9,598	11,688	+22%	1	1	0%	1,675	2,507	+50%	0	0	—
case-16 I am comparing how we fetch context for GitHub issues versus GitHub PRs when initializing worktrees. When fetching PR context for https://github.com/acme/repo/pull/200, should we use the exact same JSON field list as issue context (title, body, labels)?	fail→pass	12,919	3,941	-69%	1	1	0%	2,004	814	-59%	0	0	—
case-17 I'm done editing code inside my worktree directory for issue #104. Should I switch back to my main repository directory before running git commit, or run git commit directly in the worktree?	pass→pass	9,938	5,455	-45%	1	1	0%	1,241	1,046	-16%	0	0	—
case-18 /work Optimize database query performance for user search - I want to name this branch feature/db-perf. Is there a required prefix and date format convention for isolated worktree branches?	fail→pass	10,399	7,013	-33%	1	1	0%	1,534	1,255	-18%	0	0	—
case-19 /work https://github.com/acme/repo/pull/88 - I want to set up my worktree. Which GitHub CLI tool command and flags are used to fetch the PR metadata for this operation?	fail→fail	10,707	3,899	-64%	1	1	0%	1,674	875	-48%	0	0	—
case-20 /work https://github.com/acme/repo/issues/55 - I need to fetch issue context before starting work. What CLI command and JSON fields should be queried?	fail→fail	6,627	5,416	-18%	1	1	0%	843	846	+0%	0	0	—
case-21 I need to open a new GitHub issue for a bug found in production. Title: 'Memory leak in worker node', body: 'High RSS usage under load'. Please submit this issue to our GitHub repository.	pass→pass	8,032	5,624	-30%	1	1	0%	1,130	585	-48%	0	0	—
case-22 I have 5 unpushed commits on my current feature branch 'feature/auth-fix' and I want to combine them into a single clean commit before review. How do I perform an interactive rebase for the last 5 commits?	pass→pass	11,457	8,297	-28%	1	1	0%	1,765	1,474	-16%	0	0	—
case-23 I have uncommitted changes in my working directory that I want to temporarily save while I switch to main to test a hotfix, then restore them afterward. What git commands should I use?	pass→pass	6,650	5,569	-16%	1	1	0%	1,213	1,248	+3%	0	0	—
case-24 I have committed my changes on my local branch 'fix/login-bug' and pushed them to origin. I want to open a pull request on GitHub titled 'Fix login page authentication bypass'. What command creates the PR?	pass→pass	7,708	6,499	-16%	1	1	0%	992	917	-8%	0	0	—

These cases were also screened by the skill’s intent / safety pass.See the Trust review ↗Reproduce this benchmark in Colab ↗

How this skill was measured

DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 24 cases were attempted, and 20 counted toward the lift figure. The other 4 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +33 percentage points is the difference between those two pass rates over the 20 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.

Model measured: gemini-3.6-flash
Graded by: gemini-3.6-flash
Grading method: judged
Verification: verified
Pass rate without the skill: 25%
Pass rate with the skill: 58.3%
Cases attempted: 24
Cases counted toward lift: 20
Cases that regressed: 2
Skill version measured: v1
Run completed: 9/26/2026, 12:24 UTC

Measured example — same prompt, with vs. without the skill · from the benchmark run · case case-04

Prompt

/work https://docs.github.com/en/get-started/quickstart/hello-world - I want to research this documentation page for my upcoming task. Should we immediately create a git worktree for this web link?

Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.

Show both answers

SkillScore blends 4 signalsno signals yet

BenchmarkRun a verified benchmark with positive lift

Live evalNeeds ≥10 evaluated traces in 30 days

AI ratingNeeds ≥10 rated traces

AdoptionNeeds ≥10 serves across ≥2 orgs in 30 days

No signals yet — run a benchmark or collect evaluated traces to earn a score.

Verified on gemini-3.6-flash · re-tested yesterday

Other measured skills in the registry, with their headline benchmark lift.