Versioned security assessment

Report ID: SA-E716D641

6/30/2026, 10:56:45 AM

pexoai-agent security assessment v57

Skill Security Certification Report

Audit History
Audit model: codex Historical report
Skill name
pexoai-agent
Version
v57
Maintainer
pexoai
Source
github.com/pexoai/pexo-skills/tree/main/skills/pexo-agent/
Coverage
13 Files scanned · 2,153 Lines analyzed
Policy version
Unavailable

Highest confirmed finding severity

High

2 confirmed security findings require 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.

The static alerts are partly true positives: the skill runs bash scripts, sources a user config file, sends bearer-token API calls, uploads local media, and downloads signed assets. Many command, weak-crypto, and Windows SAM alerts are false positives from documentation, markdown examples, or ordinary shell parsing. No confirmed malicious exfiltration was found, but publication should require manual review and clear user warnings.

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

13 Files scanned · 2,153 Lines analyzed

6 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.

Observed in 3 evidence locations

Network access

May connect to external services.

Observed in 4 evidence locations

Filesystem access

May read or write local files.

Observed in 4 evidence locations

Env variables

May read values from the process environment.

Observed in 4 evidence locations

External commands

May invoke commands or programs outside the Skill.

Observed in 4 evidence locations

Capability review items (4)
High
scripts/_common.sh:7-8scripts/pexo-doctor.sh:71-76
Shell Config File Is Sourced
The shared helper sources the configured Pexo config file as shell code. A modified config can execute commands in the agent environment when any script runs.
The code directly uses source on a user-controlled config path. This is a real execution risk, although it is common for shell-based tooling.
Medium
scripts/_common.sh:32-33scripts/_common.sh:228-251
Bearer Token Sent to Configured Endpoint
API helpers build an Authorization bearer header from PEXO_API_KEY and send it to PEXO_BASE_URL. This is required for Pexo, but endpoint configuration must be trusted.
The API key header and curl calls are direct. The behavior matches the skill purpose, so this is a controlled but sensitive network pattern.
Medium
scripts/pexo-upload.sh:47-55scripts/pexo-upload.sh:83-84
Local Media Files Can Be Uploaded
The upload helper accepts a local file path and sends the file contents to a signed upload URL. This can expose sensitive files if the agent uploads the wrong path.
The data transfer is explicit and functional. It is legitimate for video generation, but it needs user consent and path review.
Medium
scripts/pexo-asset-get.sh:12-14scripts/pexo-asset-get.sh:54-60scripts/pexo-asset-get.sh:65-88
Signed Asset Downloads Written Locally
The asset helper downloads signed URLs into ~/.pexo/tmp or PEXO_TMP_DIR. This creates local copies that may persist after delivery.
The script clearly writes downloaded assets to disk. The risk depends on asset sensitivity and local file permissions.

Risk findings

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

Confirmed security concerns (2)

RISK-001 High
SKILL.md:78-86
Instruction Priority Claim in Skill Prompt
The skill text declares a language rule as highest priority and non-negotiable. This is not proven malicious, but it asserts instruction priority inside untrusted marketplace content.
The wording is explicit and high priority, but it only concerns response language and does not tell the agent to skip security checks or ignore platform policy.
RISK-002 Low
references/SETUP-CHECKLIST.md:9-15references/SETUP-CHECKLIST.md:67-75references/TROUBLESHOOTING.md:198-211
Documentation Command Examples Trigger Static Alerts
Many external command and filesystem alerts occur in setup and troubleshooting documentation. These are user-facing examples rather than hidden execution paths.
The cited lines are markdown examples and explanatory text. They do not execute unless a user deliberately follows the setup steps.

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 (1)
Low
SKILL.md:80scripts/pexo-chat.sh:10-18references/SETUP-CHECKLIST.md:89-93
Weak Crypto and Windows SAM Static Hits Are False Positives
The weak-crypto and Windows SAM detections map to ordinary words, script descriptions, or API-key prefix examples. No cryptographic operation or SAM database access was found in the cited context.
The inspected lines do not reference Windows credential files or implement weak cryptography. They appear to be scanner keyword collisions.

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