Skip to content

[WRONG BRANCH] docs(skill): keep access-key creation out of agent sessions - #373

Draft
luvs01 wants to merge 1 commit into
mainfrom
codex/propose-fix-for-agent-key-rotation-vulnerability
Draft

[WRONG BRANCH] docs(skill): keep access-key creation out of agent sessions#373
luvs01 wants to merge 1 commit into
mainfrom
codex/propose-fix-for-agent-key-rotation-vulnerability

Conversation

@luvs01

@luvs01 luvs01 commented Aug 30, 2026

Copy link
Copy Markdown
Owner

Motivation

  • An agent-facing recipe instructed models to run ocx access key create --json, which prints a one-time plaintext data-plane credential that can leak into model transcripts.
  • That disclosure allows an upstream, untrusted agent host to obtain a live admission key accepted by the server, creating an integrity/availability risk if the proxy is reachable beyond loopback.

Description

  • added a new "Secret-bearing commands" section to skills/ocx/SKILL.md that explicitly forbids running ocx access key create from an agent session and instructs the user to create keys in a separate terminal and only report the new key id.
  • updated skills/ocx/references/03_recipes.md to remove the agent-invokable ocx access key create step and to document a user-mediated rotation workflow that only removes the old key after the replacement is configured.
  • added a regression test in tests/skill-ocx.test.ts that asserts the skill contains the prohibition and that the recipes file does not include an executable ocx access key create invocation.

Testing

  • Ran the focused suite node_modules/.bin/bun test tests/skill-ocx.test.ts, which passed (12 tests, 0 failures).
  • Ran bun run typecheck and the repository bun run privacy:scan, both of which completed successfully.
  • The full bun run test suite was exercised but encountered unrelated environment/timing failures (lab sandbox / proxy-environment guards and a few timeouts); these failures are external to the documentation-only changes and do not affect the focused regression added here.

Codex Task

@coderabbitai

coderabbitai Bot commented Aug 30, 2026

Copy link
Copy Markdown

Important

  • 🔍 Trigger review

This repository does not receive automatic reviews because it has fewer than 10 stars.

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: ASSERTIVE

Plan: Pro Plus

Run ID: 46aa2e8e-9cba-4e26-b2ae-4344b2618178


Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

Comment @coderabbitai help to get the list of available commands.

@github-actions

Copy link
Copy Markdown

Deterministic PR hygiene checks passed.

@chatgpt-codex-connector

chatgpt-codex-connector Bot commented Aug 30, 2026

Copy link
Copy Markdown

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review Status Commit Review trigger
📝 Code Review Completed 2026-08-30T05:41:47.196101Z 6230c42 PR opened
ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

@github-actions github-actions Bot added the documentation Improvements or additions to documentation label Aug 30, 2026
@github-actions github-actions Bot changed the title docs(skill): keep access-key creation out of agent sessions [WRONG BRANCH] docs(skill): keep access-key creation out of agent sessions Aug 30, 2026
@github-actions

github-actions Bot commented Aug 30, 2026

Copy link
Copy Markdown

⏳ DRAFT

  • wrong target branch (main); retarget to dev.

What to do

  • Retarget this PR to dev — all contributions go to dev.

Its title has been prefixed with [WRONG BRANCH].
This pull request was already a draft. Its draft status will be preserved after every issue above is resolved.

@github-actions
github-actions Bot marked this pull request as draft August 30, 2026 05:39

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 6230c42908

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

Comment thread skills/ocx/SKILL.md

## Secret-bearing commands

**Do not run `ocx access key create`.** Its one-time response contains the new plaintext

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P1 Badge Prohibit the access-key creation alias

When an agent discovers the advertised ocx api-key create alias, this prohibition does not cover it, and src/cli/dispatch.ts forwards that alias to the same handleAccessCommand(["key", ...]) path that returns the plaintext credential. The new regression test likewise checks only the canonical spelling, so the stated secret boundary remains bypassable; prohibit all access-key creation paths from agent sessions and test both command spellings.

AGENTS.md reference: AGENTS.md:L287-L293

Useful? React with 👍 / 👎.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

aardvark codex documentation Improvements or additions to documentation

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant