v1.1.402

LobsterBio - Dev

BenignClawScan verdict for this skill. Analyzed May 1, 2026, 6:36 AM.

Analysis

The skill appears purpose-aligned for Lobster development, but it includes setup commands, optional API-key handling, and persistent session/provenance behavior that users should notice before use.

GuidanceThis looks like a normal development-oriented skill for Lobster AI. Before installing, use a project-specific virtual environment, verify the Lobster package/version you intend to install, avoid sharing real API keys in prompts or shell history, and understand that Lobster sessions/provenance may persist project context on disk.

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.

Agentic Supply Chain Vulnerabilities

SeverityLowConfidenceHighStatusNote

SKILL.md

uv venv --python 3.12 .venv && source .venv/bin/activate
uv pip install 'lobster-ai[anthropic]'

The skill directs users or agents to install an external Lobster package as part of the setup workflow. This is central to the development purpose, but it is an unpinned package install that changes the local Python environment.

User impactInstalling the package will download and execute package installation logic in the selected Python environment.

RecommendationRun setup in an isolated virtual environment, verify the package source/version, and pin versions when reproducibility matters.

Permission boundary

Checks whether tool use, credentials, dependencies, identity, account access, or inter-agent boundaries are broader than the stated purpose.

Identity and Privilege Abuse

SeverityLowConfidenceHighStatusNote

references/cli.md

--anthropic-key KEY ... --gemini-key KEY ... --openai-key KEY ... --bedrock-access-key KEY ... --bedrock-secret-key KEY

The CLI reference documents optional provider credentials and API-key configuration. These credentials are expected for LLM/provider integrations, and the artifacts do not show leakage or unrelated credential use.

User impactIf used, API keys give Lobster access to external LLM or cloud services and may be stored in local configuration files.

RecommendationPrefer environment variables or interactive setup when possible, avoid pasting real secrets into shared chats or logs, and use least-privilege/revocable keys.

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/cli.md

Sessions persist conversation history and workspace state.

The Lobster CLI documentation states that sessions can persist conversation and workspace state. This is disclosed and useful for analysis continuity, but it means project context may be reused later.

User impactPrevious analysis context, workspace state, or provenance may remain available across Lobster sessions.

RecommendationUse separate workspaces/session IDs for unrelated or sensitive projects, and review or delete session/provenance files when they should not be reused.