v1.0.9

zHive

ReviewClawScan verdict for this skill. Analyzed May 1, 2026, 5:48 AM.

Analysis

This appears to be a real zHive integration, but it needs review because it is always-on, can autonomously post predictions using a stored API key, and the registry metadata does not disclose those requirements.

GuidanceInstall only if you trust zHive and want an agent that can keep monitoring rounds and posting predictions. Treat the API key in ~/.zhive as sensitive, review the local memory/strategy files, and prefer explicit per-post approval or clear stop limits before running the autonomous loop.

Findings (5)

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

SeverityMediumConfidenceHighStatusConcern

SKILL.md

post_prediction ... Posts a price prediction to a megathread round on behalf of the agent ... prompt: per_session

The skill gives the agent authority to mutate an external zHive account by posting predictions, with approval described per session rather than per individual post.

User impactThe agent may post market predictions under the user's zHive agent identity during a session, affecting account reputation or rewards.

RecommendationOnly run it when you intentionally want posting enabled, and require explicit confirmation for each prediction or a clearly bounded session limit.

Human-Agent Trust Exploitation

SeverityMediumConfidenceHighStatusConcern

metadata

Required binaries ... none; Primary credential: none; Required config paths: none

This metadata conflicts with SKILL.md, which declares curl/jq/npx requirements, a required API key, and a plaintext config file under ~/.zhive.

User impactA user reviewing only registry metadata could underestimate the credential, filesystem, and tool access the skill expects.

RecommendationUpdate the registry metadata to disclose the required API key, config path, binaries, network domains, and plaintext storage before installation.

Rogue Agents

SeverityMediumConfidenceHighStatusConcern

SKILL.md

always: true ... After setup, connects and enters the watch loop ... enters the autonomous watch-analyze-post loop

The artifact explicitly combines persistent invocation with an autonomous loop that analyzes and posts predictions.

User impactThe skill may continue operating as a recurring agent rather than only performing a one-time user-requested action.

RecommendationDefine clear start/stop controls, maximum runtime or round limits, and visible user approval before each external posting action.

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

SeverityMediumConfidenceHighStatusNote

SKILL.md

api_key ... stored in ~/.zhive/agents/{agentName}/config.json ... Stores API key and agent name after registration. Contains plaintext config.

The zHive API key is expected for this integration, but it is a sensitive credential stored locally in plaintext.

User impactAnyone or anything that can read that local file may be able to act as the zHive agent.

RecommendationUse a dedicated zHive account/API key, restrict local file permissions, and delete or rotate the key when no longer needed.

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

SKILL.md

MEMORY.md ... Agent learnings and market observations ... Internalize these. All analysis and predictions must reflect this agent's unique voice, strategy, and biases.

The skill stores and later reuses persistent local memory and strategy files to shape future predictions.

User impactBad or stale information in those files could influence later agent behavior and posted predictions.

RecommendationReview and edit the local SOUL.md, STRATEGY.md, and MEMORY.md files periodically, especially before enabling posting.