reclaim

reclaim is a skill for Claude Code from johardi/claude-stack. It costs 157 tokens per session (1,074 once invoked), scanned A, original, MIT.

A recovery workflow for taking back a Git branch held by another coding agent's worktree. A worktree is a separate working copy connected to the same Git repository.

In plain words
What is it for?
Finding a worktree by branch or agent, saving its dirty state, releasing the branch lock, removing the worktree, and optionally checking out the branch.
Why use it?
It helps recover a branch when an agent is stuck or has stopped, while preserving uncommitted work for possible recovery.

Skill for Claude Code

Written for Claude Code: allowed-tools in frontmatter.

Runs only inside its plugin — its command needs a path that Claude Code sets for a plugin’s own hooks and for nothing else. Install the plugin, not this.

Part of the cstack-collab plugin — 5 skills, 3 hooks shipped together

Good fit Finding a worktree by branch or agent, saving its dirty state, releasing the branch lock, removing the worktree, and optionally checking out the branch.

Compare 6 skills from other repositories ↓
Install

Getting it into your agent

This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.

Claude Code
/plugin marketplace add johardi/claude-stack
Claude Code
/plugin install cstack-collab

Made for: Claude Code.

Or install cstack-collab, the plugin that ships this one along with the rest of its 5 skills, 3 hooks.

Wrote 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.

agentmods badge for reclaim

README.md
[![agentmods](https://agentmods.dev/badge/skills/johardi/claude-stack/reclaim/github.svg)](https://agentmods.dev/skills/johardi/claude-stack/reclaim)
Your own site
<a href="https://agentmods.dev/skills/johardi/claude-stack/reclaim"><img src="https://agentmods.dev/badge/skills/johardi/claude-stack/reclaim/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.

agentmods 80×15 button for reclaim

Your own site · 80×15
<a href="https://agentmods.dev/skills/johardi/claude-stack/reclaim"><img src="https://agentmods.dev/badge/skills/johardi/claude-stack/reclaim.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 157 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,074 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00157 $0.01074
Opus 5.5 $0.00063 $0.00430
Sonnet 5.5 $0.00031 $0.00215
Haiku 4.5 $0.00016 $0.00107

Measured 1mo ago against content hash ff273e416a34, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

reclaim 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 1mo 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.

plugins/cstack-collab/skills/reclaim/SKILL.md · 67 lines

How it starts

The opening of the file, as written. The whole thing — 67 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Reclaim

The complement to handoff. Where handoff is collaborative (live agent yields branch), reclaim is unilateral (human takes branch back without — or against — the agent's cooperation). Run from the primary repo, not from inside a worktree.

When to Use

Use when:

  • The user types /cstack-collab:reclaim <branch-or-agent-id> with optional flags.
  • The user describes a stuck agent worktree: "the agent crashed", "the worktree is locked but the session is dead", "I can't check out feature/X — git says it's used by another worktree".
  • An orchestrate-tickets wave left an agent worktree behind that needs to be cleaned up before the next wave.

Skip when:

  • The agent is alive and reachable. Tell the user to type /cstack-collab:handoff in the agent's session instead — handoff preserves the agent's context and lets it write a meaningful commit message. Reclaim is the rescue tool, not the default.
  • The user just wants to list what's stuck — that's /cstack-collab:list-worktrees.

Inputs

Argument Meaning
<ref> (positional, required) Branch name (feature/66-x), agent id (agent-a81ac7c0…), or absolute path to the worktree.
--force Remove even when the agent's PID is still alive. Kills the live agent's session. Only pass when the user has explicitly acknowledged this — e.g., "the agent is wedged, just kill it".
--stash (default) preserve dirty state at refs/cstack-collab/reclaim/<agent-id>.
--discard Throw the dirty state away.
--checkout After removal, run git checkout <branch> in the primary repo.

How to Run

python3 "${CLAUDE_PLUGIN_ROOT}/scripts/reclaim.py" <ref> [flags]

The script:

  1. Locates the agent worktree. Refuses if the ref doesn't match exactly one.
  2. Refuses if the agent's PID is alive, unless --force was passed.
  3. Stashes any dirty state to refs/cstack-collab/reclaim/<id> (unless --discard). The ref points at a real commit object, so it survives git worktree remove — recover with git stash apply <ref> or git cherry-pick <ref>.
  4. git worktree unlock + git worktree remove [--force].
  5. Optional git checkout <branch> in the primary repo.
  6. Appends a reclaimed event to state.json if an orchestration is active.

Read the full file on GitHub · 67 lines

Changes

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.

  1. 1mo ago First seen · 67 lines · 157 tokens per session scan A ff273e416a34

Subscribe to this mod's changes

reclaim is a skill published in the GitHub repository johardi/claude-stack (2 stars, last pushed 5mo ago), licensed MIT. It adds 157 tokens to every session and 1,074 once invoked, about $0.0006 per session on Opus 5.5. 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.