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 skills add hams-ollo/zen-agent-skills --skill reconcile-worktreesgit clone --depth 1 https://github.com/hams-ollo/zen-agent-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/hams-ollo/zen-agent-skills/reconcile-worktrees)<a href="https://agentmods.dev/skills/hams-ollo/zen-agent-skills/reconcile-worktrees"><img src="https://agentmods.dev/badge/skills/hams-ollo/zen-agent-skills/reconcile-worktrees/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/hams-ollo/zen-agent-skills/reconcile-worktrees"><img src="https://agentmods.dev/badge/skills/hams-ollo/zen-agent-skills/reconcile-worktrees.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What 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 5.1 | $0.00161 | $0.03766 |
| Opus 5 | $0.00081 | $0.01883 |
| Sonnet 5 | $0.00032 | $0.00753 |
| Haiku 4.5 | $0.00016 | $0.00377 |
Grade A, and why
reconcile-worktrees 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 6d 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 — 263 lines — stays where its author put it; the contents beside it link to each section on GitHub.
reconcile-worktrees
Bring the verified contents of one or more isolated git worktrees into the main working tree as a clean, reviewable diff, without silently overwriting anything, without committing on the user's behalf, and without trusting that "no conflicts reported" means "safe".
This is the landing step of the kit spine: new-task authors the task
files, fix-batch dispatches them to isolated worktree agents, and this
skill consolidates the results. It assumes each worktree has already been through its own
verification pass via verifier-agent (fix-batch's Step 6 is what
runs it against every worktree in a batch). This skill is about combining already-trusted changes,
not about trusting them in the first place. If a worktree has not been independently verified yet,
run verifier-agent against it first. The same holds for the report
the worktree arrived with: it meets fix-batch's delegate report contract, or the worktree does not
land. Step 1 is where that is checked.
Landing the diff is where this skill stops, and it is not where the spine stops. A batch that has
just been consolidated is the single change most likely to have invalidated the reader-facing
documents, and nothing above notices, so doc-sync runs next, over the combined result rather than
over any one worktree. Offer it as the natural next step when you report in Step 7, and say so even
when the diff looks purely internal, since that judgment is what doc-sync exists to make.
Why this exists
Reconciling several worktrees by hand after a batch of parallel agent fixes turned out to be its
own source of risk, separate from whatever the agents themselves did wrong. In the 2026-07-07
incident that motivated the fix-batch skill: task-file bookkeeping split three incompatible ways
across worktrees because some agents wrote directly to the main checkout and others did not; a
binary asset showed an identical phantom diff in every worktree due to an unmigrated git-lfs rule,
which would have silently corrupted the file if any worktree's changes were committed with a
blanket git add -A; and two worktrees both touched the same file in different, non-overlapping
regions, which is safe to combine but only if you actually check that before assuming it.
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.
- 6d ago Changed · +7 lines bbd26aa19de8
- 10d ago First seen · 256 lines · 161 tokens per session scan A 12c5a6bfc7d4
reconcile-worktrees is a skill published in the GitHub repository hams-ollo/zen-agent-skills (2 stars, last pushed 8d ago), licensed MIT. It adds 161 tokens to every session and 3,766 once invoked, about $0.0008 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 2026-08-31.
Other skills, from other repositories
rig-epic
Plan and run a multi-ticket epic: decompose a feature into parent + child items and stack PRs on a shared integration branch instead of landing each on main. Use when children interleave (one item's runtime contract depends on another's incomplete state) — stacking keeps each child PR reviewable without temporarily…
approve-proposed
Atomically persist every file marked PROPOSED in the most recent prior assistant turn's ### Artifact changes block. Single-command idiom that closes the two-step latency in ADR-0001's PROPOSED-by-default contract; the user reviews proposals in Ask/Plan mode, then runs this once to write all of them. Use when the prior…
rig-worktree
Manage isolated git worktrees: create one wired up for dev (fetch, branch from base, symlink env, install deps), list them with PR state, or safely remove merged ones. A shared bootstrap that implement-style flows can call, also usable directly. Triggers on: 'worktree', 'set up a worktree', 'new worktree', 'list…
rig-task
Implement one unit of work end-to-end — from a tracker issue OR an ad-hoc description: spec review, TDD (RED→GREEN→REFACTOR), pre-PR self-review, open a PR, then drive the review-bot loop to clean. Runs start→finish in one shot by default; start/finish are optional phases for pause/resume. Sibling to /rig-epic (one…
pr-package
Prepare a clean delivery package for GitHub from the real git diff vs an explicit base branch, persisted as PRPACKAGE.md (or PRPACKAGE. .md for multi-repo tasks). Produces branch name, commit messages, fetch, checkout, add, commit, and push commands, PR title and body, and reviewer attention points; never invents work…
branch-commit
Return a branch name and a concise commit message (at most 2 lines) for the current task, grounded in the real git diff rather than a paraphrase of the task summary, and with --apply optionally create that commit after showing it and getting the user's confirmation in the same turn. Use when the user only needs quick…