Agent Identity

ReviewAudited by ClawScan on May 10, 2026.

Overview

The skill mostly matches its cryptographic identity purpose, but its key-generation code can write files outside the intended keys folder if given a crafted agent name.

Use this skill only with simple, trusted agent names until path validation is fixed; do not use names containing slashes, backslashes, drive letters, or '..'. Protect generated private keys, avoid real passwords on the command line, and install dependencies from a trusted source.

Findings (4)

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

A crafted agent name could cause the skill to create or overwrite key files outside the expected folder.

Why it was flagged

The required --name value is used directly in a filesystem path and later opened for writing, so names containing absolute paths or parent-directory segments could escape the intended keys/ directory.

Skill content
private_path = os.path.join(KEY_DIR, f"{name}_private.pem")
Recommendation

Restrict agent names to safe characters, reject path separators and absolute paths, and resolve/validate that outputs remain inside the intended keys directory before writing.

What this means

Anyone who obtains the private key, or its password if encrypted, could sign messages as that agent identity.

Why it was flagged

The skill intentionally creates and uses private keys for agent identity, which is purpose-aligned but sensitive.

Skill content
Keys are stored in `keys/` directory ... Never share your private key
Recommendation

Use encryption for private keys, avoid command-line passwords for real use, store keys with restrictive file permissions, and back them up securely.

What this means

The installed package version depends on the user’s package index and environment.

Why it was flagged

The Python dependency is expected for cryptographic operations, but the install instruction is unpinned and there is no separate install spec or lockfile in the supplied artifacts.

Skill content
pip install cryptography
Recommendation

Install from a trusted package index and consider pinning a known-good cryptography version in a controlled environment.

What this means

Windows instructions may not work as written, and users should not substitute an unreviewed wrapper from elsewhere without checking it.

Why it was flagged

The documentation references a PowerShell wrapper, but the provided file manifest only includes SKILL.md and identity.py, so the recommended Windows wrapper is absent from the reviewed artifacts.

Skill content
`agent-identity.ps1` - PowerShell wrapper (Windows)
Recommendation

Use the reviewed Python CLI directly or obtain any missing wrapper only from a trusted, reviewed source.