manage-task-checkpoints

manage-task-checkpoints is a skill for Claude Code, Codex from alibaba/anolisa. It costs 67 tokens per session (937 once invoked), scanned A, original, Apache-2.0.

A recovery workflow for snapshots of COSH-managed Tasks. A snapshot is a saved workspace state that can be inspected, compared with the current files, or restored.

In plain words
What is it for?
Use it to list Task snapshots, preview or compare a selected snapshot, switch a Task workspace, and retry after a workspace-only failure.
Why use it?
It helps recover a Task workspace after a file-related failure without using the wrong workspace or recovery point. It also keeps recovery limited to the managed Task it belongs to.

Skill for Claude CodeCodex ✓ vendor

Written for no agent in particular: nothing here depends on one.

Good fit Use it to list Task snapshots, preview or compare a selected snapshot, switch a Task workspace, and retry after a workspace-only failure.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/alibaba/anolisa/manage-task-checkpoints
About the project

ANOLISA is a server-side operating layer for AI agent workloads that provides terminal access, token-saving tool-output compression, runtime controls, security, observability, skills, memory, and sandbox management. It is for running and supervising agents from a Linux terminal while retaining an existing shell, agent framework, and sandbox. The catalogue add-ons are components of its agent operating environment and workflows.

alibaba/anolisa · 603 stars · on GitHub · agentic-os.sh

Install

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.

Any agent
npx skills add alibaba/anolisa --skill manage-task-checkpoints
Clone the repo
git clone --depth 1 https://github.com/alibaba/anolisa

Made for: Claude Code, Codex.

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 manage-task-checkpoints

README.md
[![agentmods](https://agentmods.dev/badge/skills/alibaba/anolisa/manage-task-checkpoints/github.svg)](https://agentmods.dev/skills/alibaba/anolisa/manage-task-checkpoints)
Your own site
<a href="https://agentmods.dev/skills/alibaba/anolisa/manage-task-checkpoints"><img src="https://agentmods.dev/badge/skills/alibaba/anolisa/manage-task-checkpoints/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 manage-task-checkpoints

Your own site · 80×15
<a href="https://agentmods.dev/skills/alibaba/anolisa/manage-task-checkpoints"><img src="https://agentmods.dev/badge/skills/alibaba/anolisa/manage-task-checkpoints.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 67 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 937 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 $0.00067 $0.00937
Opus 5 $0.00034 $0.00468
Sonnet 5 $0.00013 $0.00187
Haiku 4.5 $0.00007 $0.00094

Measured today against content hash 04ec8794d5b5, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade A, and why

manage-task-checkpoints 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 today.

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.

src/cosh-ng/skills/manage-task-checkpoints/SKILL.md · 98 lines

How it starts

The opening of the file, as written. The whole thing — 98 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Manage COSH Task Checkpoints

Treat this Skill as a recovery workflow, not a security boundary. Gateway and ws-ckpt enforce Task ownership, complete snapshot IDs, workspace generation, live diff, peer identity, and replay safety.

Keep the scope Task-only

  • Use this Skill only for /task, tsk_..., or an explicitly managed Task.
  • Use the separate ws-ckpt Skill for a standalone workspace.
  • Do not infer the managed workspace from cwd or ask the user for its path.
  • Do not call bare ws-ckpt, a lab script, or the ws-ckpt socket for a Task.

Inspect Task snapshots

Use the Task-scoped Gateway CLI. It resolves an explicit socket, then COSH_GATEWAY_SOCKET, then the active packaged Gateway instance:

cosh-gateway task --output jsonl list
cosh-gateway task --output jsonl snapshot list <task-id>
cosh-gateway task --output jsonl snapshot preview <task-id> <snapshot-id>
cosh-gateway task --output jsonl snapshot diff <task-id> <snapshot-id>

Accept only complete Task and snapshot IDs returned by Gateway. Never guess or expand a prefix. When several Tasks or recovery points match, summarize their status and file effects and ask the user to choose by meaning rather than copying internal IDs.

Users can inspect the same state directly with:

/task snapshots <task-id>
/task snapshot preview <task-id> <snapshot-id>
/task snapshot diff <task-id> <snapshot-id>

Switch safely

  1. Require the Task to be terminal before switching.
  2. Ensure the outer cosh-shell process was launched outside the managed workspace. Running cd /tmp inside an Agent Bash command changes only the child process and does not release the parent shell cwd.
  3. Preview the exact snapshot immediately before switching.
  4. Describe the files that will be removed, restored, or retained. Ask for natural-language confirmation unless the user already authorized this one exact recovery attempt.
  5. Submit the exact snapshot ID, preview digest, terminal Task revision, and a fresh idempotency key for this operation.
  6. Verify the resulting file state against the preview.

Read the full file on GitHub · 98 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. today First seen · 98 lines · 67 tokens per session scan A 04ec8794d5b5

Subscribe to this mod's changes

manage-task-checkpoints is a skill published in the GitHub repository alibaba/anolisa (603 stars, last pushed today), licensed Apache-2.0. It adds 67 tokens to every session and 937 once invoked, about $0.0003 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 2026-09-10.