Versioned security assessment

Report ID: SA-706BD8AA

6/30/2026, 8:28:33 AM

lobster-cognitive-growth security assessment v2

Skill Security Certification Report

Audit History
Audit model: codex Historical report
Skill name
lobster-cognitive-growth
Version
v2
Maintainer
norika1207-lab
Source
github.com/norika1207-lab/lobster-cognitive-growth-skill/tree/main/
Coverage
2 Files scanned · 331 Lines analyzed
Policy version
Unavailable

Highest confirmed finding severity

Low

1 confirmed security finding requires attention.

Installation context

Historical evidence

This report may not describe the currently installable artifact. Open the current Skill page for install guidance.

Open current Skill page

This report does not block or authorize the manifest or ZIP.

Static analysis reported many command, network, reconnaissance, and weak-cryptography patterns, but most command and cryptography hits are Markdown false positives. The confirmed concerns are an optional authenticated Charenix integration, an unpinned remote curl install command, and external links that require user trust; no confirmed malicious intent or automatic exfiltration was found.

Report position

Historical report

Open audit history before using this report to install.

Audit attestation

Not attestable

The required immutable binding is incomplete.

Human verification

Not verified

No human verification is recorded for this report.

Coverage

2 Files scanned · 331 Lines analyzed

4 items shown for review

Limitations

This report does not claim runtime or sandbox execution and does not prove the absence of side effects.

Evidence chain

Follow the evidence from source binding to the install contract. Available evidence supports verification; it is not a safety guarantee.

  1. Source

    Binding unavailable

  2. Artifact

    Identity incomplete

  3. Audit

    Complete

  4. Install contract

    Open manifest to verify

    Open manifest

Capabilities observed

Observed means this report recorded supporting evidence. Not recorded does not prove that a capability is absent.

Contains scripts

May execute code included with the Skill.

Not recorded by this audit

Network access

May connect to external services.

Observed in 4 evidence locations

Filesystem access

May read or write local files.

Not recorded by this audit

Env variables

May read values from the process environment.

Observed in 2 evidence locations

External commands

May invoke commands or programs outside the Skill.

Observed in 1 evidence location

Capability review items (3)
Medium
SKILL.md:233-256
Authenticated External Persistence Integration
The skill documents Charenix API endpoints as a persistence layer and instructs users to send X-Agent-Key on protected requests. This is a legitimate integration pattern, but journals, history, strategies, and an agent key may leave the local environment if users enable it.
The endpoint list, base URL, and authentication header are explicit. I did not find evidence that the skill automatically sends data, so the concern is controlled external data sharing rather than confirmed exfiltration.
Medium
README.md:11-14
Unpinned Remote Install Command
The README recommends a curl command that downloads SKILL.md from a branch URL into the local skills directory. This is common for installation docs, but the target can change over time because it is not pinned to an immutable commit.
The shell command and remote URL are directly present in the README. The risk depends on a user manually running the command, and no auto-execution code was found.
Low
README.md:28-30SKILL.md:5-7
External Documentation Links Require User Trust
The skill and README include external links to Charenix, GitHub, ClawHub, and AgentSkill. These are documentation and listing links, not automatic network requests, but users should verify the destinations before using integrations.
The URLs are plainly visible and are not wrapped in executable code at these locations. The residual risk is user navigation or optional integration, not hidden network activity.

Risk findings

Confirmed security concerns are separated from items that still need review.

Confirmed security concerns (1)

RISK-001 Low
SKILL.md:259-279
Agent-Facing Owner Prompt Should Remain Untrusted Content
The skill includes an owner prompt that tells the agent it is allowed to use the skill and sets goals for behavior. I did not find instructions to ignore higher-priority rules, skip review, or override system policy, so this is not confirmed prompt injection.
The block is clearly a prompt template and uses agent-directed wording. It lacks the stronger override language that would justify a high or critical prompt-injection finding.

Expert evidence

Immutable subject identity, scanner metadata, dismissed matches, and source-level evidence.

Artifact subject

Marketplace commit
Unavailable
Content hash
Unavailable
Tree hash
Unavailable
Skill path
Unavailable
Audit payload hash
Unavailable

Analysis metadata

Audit model: codex

Analysis state: Complete

Scope is limited to the recorded files, lines, methods, and evidence. No runtime or sandbox execution is claimed.

Static false positives ignored (2)
Low
SKILL.md:69-73SKILL.md:172-190SKILL.md:222-231
Static Backtick Command Findings Are Markdown False Positives
Most Ruby or shell backtick detections occur inside fenced text, HTTP, JSON, or prompt examples. They are documentation examples, not executable code paths inside this skill.
The cited locations are fenced Markdown examples and contain no runnable script files or invocation mechanism. This strongly supports dismissing the bulk of the command-execution detections as false positives.
Low
SKILL.md:3SKILL.md:18
Weak Cryptography Static Findings Are False Positives
The weak cryptography detections map to prose describing intelligence and cognition, not to MD5, SHA1, DES, RC4, or similar cryptographic use. No cryptographic operation was found in the cited lines.
The cited lines are plain descriptive text about agent training. There is no code, algorithm name, digest call, or security-sensitive cryptographic context.

Verify and export

JSON report CITATION.cff SARIF 2.1.0 Signed installation manifest

The manifest and lockfile bind install artifacts to cryptographic hashes. This integrity claim is separate from the security assessment.

Audit attestation: not_attestable