# Path traversal and arbitrary directory deletion/overwrite via agent profile import in local-operator

Advisory record · `GHSA-22mg-8gw7-636x`

- **Identifiers:** `GHSA-22mg-8gw7-636x` · `PYSEC-2026-4009`
- **State:** Confirmed
- **Severity:** CVSS 9.1 Critical
- **Critical:** Yes: CVSS 9+
- **Weakness (CWE):** [CWE-22](https://cwe.mitre.org/data/definitions/22.html)
- **Tags:** [Path traversal](https://agentprime.io/advisories/tag/path-traversal.md)
- **Published:** 5 Sept 2026
- **Updated:** 30 Sept 2026
- **Reviewed by a person:** No

Published by the GitHub Advisory Database at 5 Sept 2026, 07:43 UTC. agentprime has not analysed this yet, so whether it affects any app is not known.

## Critical

Critical, because:

- CVSS score 9.1: 9.0 or more.

How we decide what’s critical: <https://agentprime.io/advisories/critical.md>

## Affected packages

- `local-operator` (PyPI): affected `<29fc4c68a3ed1d238b8abf6b6a6136a3361416a5 || <0.47.5`; fixed in `29fc4c68a3ed1d238b8abf6b6a6136a3361416a5`, `0.47.5`

## Machine-written summary

Summary written by Claude Haiku from the source advisory. Not reviewed by a person.

> Local-operator versions before 0.47.5 let the agent profile import endpoint (/v1/agents/import) and AgentRegistry.import_agent() build destination directories from the id field in an uploaded agent.yml. An unauthenticated client with network access to the API can use a crafted id to recursively delete arbitrary directories and write files outside the registry, with the server process's privileges. Upgrade to 0.47.5, which assigns server-generated identifiers to imported profiles and binds lop serve to 127.0.0.1 by default.

Written 1 Nov 2026.

## Not analysed yet

We haven’t published our analysis of this Advisory yet. Once a person has reviewed it, this page will say what’s vulnerable, what has to be true for your app to be affected, and how to check.

## Severity and exploitation

- 9.1 Critical, CVSS 3.1: `CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:H/A:H`, from OSV

- **CISA KEV:** Not listed.
- **EPSS:** No score.

## Timeline

- 31 Oct 2026, 14:09 UTC: First seen by agentprime
- 31 Oct 2026, 14:09 UTC: [Confirmed by OSV](https://osv.dev/vulnerability/PYSEC-2026-4009)

## Sources

- [GitHub Advisory Database: GHSA-22mg-8gw7-636x](https://github.com/advisories/GHSA-22mg-8gw7-636x)
- [OSV: GHSA-22mg-8gw7-636x](https://osv.dev/vulnerability/GHSA-22mg-8gw7-636x)
- <https://github.com/damianvtran/local-operator/security/advisories/GHSA-22mg-8gw7-636x>
- <https://github.com/damianvtran/local-operator/pull/643>
- <https://github.com/damianvtran/local-operator/commit/29fc4c68a3ed1d238b8abf6b6a6136a3361416a5>
- <https://github.com/damianvtran/local-operator/releases/tag/v0.47.5>
- <https://pypi.org/project/local-operator/>

Facts on this page come from the GitHub Advisory Database (CC-BY 4.0), OSV and the other sources listed, as published. We don’t reproduce full advisory descriptions, which can contain exploit code.

## Check whether this affects you

A Match means an installed version is in the affected range above. Most don’t affect the app; agentprime hasn’t analysed this one yet.

- With the agentprime MCP server (<https://api.agentprime.io/mcp>), ask your agent to check the project: `check_lockfile` sends a lockfile, `check_packages` a list of installed packages, and both return the Matches and the upgrades that clear them. `get_advisory` returns this record. Setup: <https://agentprime.io/mcp.md>
- In a browser: <https://agentprime.io/check/> reads your lockfile in the browser, never uploading it.
- Whether it touches your code, with file and line: the Claude Code plugin, <https://agentprime.io/plugin.md>
- As JSON: <https://api.agentprime.io/v1/advisories/GHSA-22mg-8gw7-636x>, or `POST https://api.agentprime.io/v1/match` with your packages (<https://agentprime.io/api.md>)

---

The web page: <https://agentprime.io/advisories/GHSA-22mg-8gw7-636x/>. Everything for agents: <https://agentprime.io/llms.txt>.