Codexmonitor

PassAudited by VirusTotal on May 13, 2026.

Overview

Type: OpenClaw Skill Name: codexmonitor Version: 0.2.2 The skill bundle is designed to install and use `codexmonitor` via Homebrew to inspect local OpenAI Codex sessions. All instructions in `SKILL.md` and `SETUP.md` are straightforward, focusing on installation and usage of the `codexmonitor` tool. There is no evidence of prompt injection against the agent, data exfiltration, malicious execution, persistence mechanisms, or obfuscation. The access to `~/.codex/sessions/` is explicitly stated and directly aligns with the skill's stated purpose of monitoring Codex sessions.

Findings (0)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

Using the skill may bring private Codex session contents into the assistant conversation or command output.

Why it was flagged

The skill intentionally reads persistent local Codex session history, which may include sensitive prompts, code, outputs, or other private context.

Skill content
Reads sessions from ~/.codex/sessions by default (or via CODEX_SESSIONS_DIR / CODEX_HOME overrides).
Recommendation

Use it only when you intend to inspect those sessions, prefer specific session IDs or scoped directories, and avoid sharing outputs that contain secrets or private code.

What this means

Installing the skill means trusting the Homebrew formula and binary provider for local file-reading behavior.

Why it was flagged

The skill depends on installing and running a binary from an external Homebrew tap rather than code included in the skill artifacts.

Skill content
brew tap cocoanetics/tap
brew install codexmonitor
Recommendation

Review the Homebrew formula or project repository before installing, and install only from a tap you trust.

What this means

If used unintentionally, it could modify an existing Codex session rather than merely inspect it.

Why it was flagged

This documented command is outside pure read-only monitoring and can append to or resume an existing Codex session if used.

Skill content
Sessions can be resumed/appended by id via Codex: `codex exec resume <SESSION_ID> "message"`.
Recommendation

Use the resume command only when the user explicitly asks to continue a Codex session; otherwise keep interactions to list/show/watch inspection commands.