Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add commands/davila7/claude-code-templates/worktree-delivergit clone --depth 1 https://github.com/davila7/claude-code-templatesWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00012 | $0.01000 |
| Opus 5 | $0.00006 | $0.00500 |
| Sonnet 5 | $0.00002 | $0.00200 |
| Haiku 4.5 | $0.00001 | $0.00100 |
Grade A, and why
worktree-deliver scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 128 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Worktree Deliver
Commit all work, push, and create a pull request from the current worktree.
Instructions
You are inside a worktree. Package up the work and deliver it as a PR.
Step 1: Validate Environment
- Verify this is a worktree (not the main working tree) using
git worktree list - Get current branch:
git branch --show-current - Verify branch follows
claude/*,claude-daniel/*, orreview/*pattern. If not, warn the user and ask if they want to continue. - Read
.worktree-task.mdif it exists to get the original task description
Step 2: Review Changes
- Run
git diff --statandgit diff --cached --statto show all changes - Run
git status --shortto show the full picture - If there are no changes at all (clean working tree, no commits ahead of main), inform the user there's nothing to deliver and stop.
Step 3: Clean Up Task File
Before staging anything, remove the worktree task file so it doesn't end up in the commit:
rm -f .worktree-task.md
Step 4: Confirm Files to Commit
Use AskUserQuestion to show the user what will be committed and ask for confirmation. List all modified, added, and untracked files.
Options:
- "Stage all changes" — stage everything
- "Let me choose" — user will specify which files to include
If the user wants to choose, ask them which files to stage.
Step 5: Stage and Commit
- Stage the confirmed files with
git add - Generate a commit message following conventional commits format
Commit Message Strategy:
-
Analyze the diff to determine the conventional commit type:
feat:— New functionality, new files, new exports, new API endpointsfix:— Bug fixes, error corrections, fixing broken behaviorrefactor:— Code restructuring without changing behaviordocs:— Documentation only changestest:— Adding or modifying testschore:— Build scripts, configs, maintenance tasks
-
Generate a commit message based on:
- The task description from
.worktree-task.md(if it was found) - A brief summary of what the diff actually changed
- Format:
<type>: <subject>(max 72 characters)
- The task description from
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 128 lines · 12 tokens per session scan A ec3a16c296b8
worktree-deliver is a command published in the GitHub repository davila7/claude-code-templates (30,476 stars, last pushed today), licensed MIT. It adds 12 tokens to every session and 1,000 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other commands, from other repositories
README
Git workflow and quality assurance commands for the claude-skills repository.
become
Agent crystallization command. Studies a tool, role, or person and produces a portable 9-section agent that inhabits the domain - with beliefs, scar tissue, and instincts.
cycle-start
Activate a cycle and start implementing its stories.
research-verify
Verify existing research findings against independent primary sources. Upgrades confidence from 'sources agree' to 'independently verified.'.
story-continue
Resume an interrupted story from where you left off.
story-archive
Move a story from a cycle back to the backlog.