Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
Codex Profiler
v1.0.27Manage OpenAI Codex profiles via Telegram commands for usage checks and OAuth token refresh with queued safe apply and progress updates.
⭐ 0· 511·1 current·1 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
Name/description align with the included scripts: usage checks against chatgpt.com and OAuth helper for Codex profiles in ~/.openclaw. However, SKILL.md repeatedly states gateway-native commands should be preferred and that auth files should not be hand-edited; the shipped scripts nevertheless directly write/overwrite auth-profiles.json and openclaw.json, and schedule on-disk shell scripts to copy/replace those files—an inconsistency between claimed approach and actual capability.
Instruction Scope
SKILL.md claims 'No remote shell execution' and 'Never mutate auth-profiles.json directly in normal operations', yet the scripts (codex_auth.py and codex_usage.py) read and write ~/.openclaw/*.json, create backups, build and execute shell scripts that cp files, run pkill, and restart/open openclaw-gateway via nohup/systemd-run. Those actions are outside the narrow scope of simple read-only usage checks and OAuth status queries described, and they perform system-level mutations without the SKILL.md appearing to require explicit user confirmation for all cases.
Install Mechanism
There is no install spec (instruction-only skill) which lowers supply-chain risk, but the bundle includes substantial Python scripts that will execute local system commands when run. Because code will be executed directly, lack of an install step does not imply lack of side effects—the code itself performs file/system changes.
Credentials
The skill requests no external credentials and operates on local OpenClaw config/auth files in the user's home directory and /tmp. Access to ~/.openclaw/auth-profiles.json and the ability to restart or spawn the openclaw-gateway are expected for a profile management tool, but these are high-privilege actions (can replace auth data and restart gateway) and exceed mere read-only checks. No unrelated environment variables or external credentials are requested.
Persistence & Privilege
The skill does not request always:true, but the code writes backups and staged auth files under /tmp, spawns off-host scripts (systemd-run / nohup), and forcibly kills/starts gateway processes (pkill, openclaw-gateway). Those actions grant the skill effective system-level influence while running and increase the blast radius if invoked autonomously. Consider requiring explicit manual approval for apply/delete operations.
What to consider before installing
This package broadly does what it says (usage checks and OAuth helper), but there are important inconsistencies and risky behaviors you should consider before installing or running it:
- Inconsistency: SKILL.md says prefer gateway-native commands and avoid hand-editing auth files, yet the included scripts do direct writes to ~/.openclaw/auth-profiles.json and OPENCLAW config and create/execute shell scripts that copy/replace those files and restart the gateway.
- System effects: The scripts call pkill, spawn openclaw-gateway via nohup or systemd-run, and write staged files in /tmp. These are system-level operations that can affect your running gateway and auth state—backup first.
- Review & test: Inspect the full scripts yourself (or with a trusted engineer). If you plan to run mutation/apply actions, test in a non-production environment or on a throwaway agent account first.
- Operational safeguards: Require manual confirmation before any 'apply' or 'hard delete' action, keep backups of ~/.openclaw/*, and prefer running usage-only/read-only commands initially. Consider disabling autonomous invocation for mutation paths or enforcing a human-in-the-loop confirmation (especially for deletes or gateway restarts).
- Ask the author: Request clarification on why scripts perform direct file writes and gateway restarts despite the SKILL.md guidance; ask for a mode that strictly uses only the gateway-native CLI commands without direct file mutation.
- Note about code quality: There are signs of sloppy/unfinished code (a truncated/misspelled token in the gateway-start section) that may cause runtime errors; this suggests the code may not have been thoroughly tested in your environment.
If you do not trust the author or cannot safely test in a sandbox, avoid granting this skill the ability to perform apply/delete operations.Like a lobster shell, security has layers — review code before you run it.
latestvk97c9athfmfct5nkr8t3wd2kps835k50
License
MIT-0
Free to use, modify, and redistribute. No attribution required.
