v0.1.0

MessageGuard

BenignClawScan verdict for this skill. Analyzed May 1, 2026, 6:01 AM.

Analysis

MessageGuard appears to be a coherent protective outgoing-message filter, with disclosed caveats around message interception, optional local logging, and verifying the external clone source.

GuidanceThis skill looks appropriate for preventing accidental secret leaks, but review its configuration before use. Pay special attention to whether logging is enabled, where logs are stored, which channels or patterns bypass filtering, and whether the GitHub repository you clone matches the reviewed code.

Findings (3)

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.

Abnormal behavior control

Checks for instructions or behavior that redirect the agent, misuse tools, execute unexpected code, cascade across systems, exploit user trust, or continue outside the intended task.

Tool Misuse and Exploitation

SeverityLowConfidenceHighStatusNote

references/integration.md

The filter is a standalone Python script that acts as a **gate** between the agent's intent to send a message and the actual `message` tool call.

The skill is designed to control the outbound message path and may mask, warn, or block messages before sending. This is the stated protective purpose, but it affects user-visible communication.

User impactOutgoing messages may be changed or stopped based on regex matches and configuration.

RecommendationReview the configured mode, built-in patterns, allow-lists, and exit-code handling before integrating it into any automatic send workflow.

Agentic Supply Chain Vulnerabilities

SeverityLowConfidenceMediumStatusNote

SKILL.md

Clone the repository: `git clone git@github.com:AndrewAndrewsen/MessageGuard.git`.

The installation guidance points users to an external GitHub repository rather than a declared install spec. The clone is user-directed, but users should verify the repository provenance before running the script.

User impactA user could install code from a repository they have not verified.

RecommendationConfirm the repository owner, commit, and contents match the reviewed artifact before using the cloned code.

Sensitive data protection

Checks for exposed credentials, poisoned memory or context, unclear communication boundaries, or sensitive data that could leave the user's control.

Memory and Context Poisoning

SeverityLowConfidenceHighStatusNote

references/integration.md

Detection log format (`log_path`, JSONL): `{"ts": "2026-02-24T13:00:00Z", "channel": "discord_general", "blocked": true, "detections": [{"name": "jwt_token", "action": "block", "snippet": "eyJhbGc…"

If logging is enabled, the skill persists channel information and snippets of detected sensitive values. This is disclosed and optional, but the log can still contain sensitive context.

User impactLocal log files may reveal where sensitive data was detected and partial prefixes of secrets or personal data.

RecommendationKeep detection logging disabled unless needed; if enabled, store logs in a protected location, restrict file permissions, and rotate or delete logs when no longer needed.