ClawHub
Back to skill

Security audit

Agentic Discovery Skill

Security checks for vulnerabilities and agentic risk

Overview

This is a local CompleteTech discovery-document skill with disclosed branded PDF rendering, not evidence of credential theft, persistence, or exfiltration.

Install only if CompleteTech-branded discovery artifacts are acceptable for your use case. Review generated PDFs before sending them externally, override footer/watermark values where appropriate, and consider updating or disabling optional PNG preview dependencies if your environment enforces strict dependency hygiene.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Findings (1)

T01 · Skill Instruction Hijacking

Warning
Location
SKILL.md:100
Finding

Mandatory Third-Party Branding and Contact Promotion in Generated Documents

Content
View full analysis

Vulnerability Details

File Location: SKILL.md:100-105; scripts/render_discovery.py:75-76, 134-135; scripts/render_pdf.py:51-55, 371-375
Vulnerability Type: Output instruction hijacking through mandatory branding and hardcoded promotional content
Risk Level: Medium

Complete Code Snippets

SKILL.md:100-105:

markdown
Artifacts from this skill are delivered as branded CompleteTech LLC **PDF** documents, not raw Markdown. The renderer emits the PDF (and prints the Markdown) in **one command**, using the same reportlab branding engine as the contract skill:

```bash
pip install -r requirements.txt
python3 scripts/render_discovery.py --template requirements-brief-for-proposal-sow-handoff \

scripts/render_discovery.py:75-76:

python
parser.add_argument("--watermark", default="DEMO DRAFT", help="Watermark text (empty to disable).")
parser.add_argument("--footer", default="CompleteTech LLC - demonstration artifact", help="Footer text.")

scripts/render_discovery.py:134-135:

python
"watermark": args.watermark, "footer": args.footer,
"disclaimer": "Demonstration artifact — replace placeholders with verified facts before use.",

scripts/render_pdf.py:51-55:

python
"zebra": "#F8FAFC",
"brand_name": "CompleteTech",
"brand_tagline": "Innovation at Every Integration",
"contact": "complete.tech  ·  Timothy.Gregg@complete.tech",
}

scripts/render_pdf.py:371-375:

python
c.setFont("Helvetica", 8.5)
c.setFillColor(colors.Color(1, 1, 1, alpha=0.78))
c.drawString(nx, h - 0.58 * inch, BRAND["brand_tagline"])
c.setFont("Helvetica", 7.5)
c.setFillColor(colors.Color(1, 1, 1, alpha=0.62))
c.drawString(nx, h - 0.74 * inch, BRAND["contact"])

Technical Analysis

The skill instructions state that artifacts are to be delivered as CompleteTech-branded PDF documents. The associated renderer reinforces that instruction by supplying company- ...[truncated 2169 chars]

Remediation
View remediation

Remediation Suggestions

  1. Make branding explicitly opt-in rather than mandatory in SKILL.md.
  2. Replace company-specific defaults with neutral values, such as an empty footer, no contact address, and no promotional tagline.
  3. Accept branding through explicit command-line options or a user-provided configuration file.
  4. Do not render a website, email address, company identity, logo, watermark, or footer unless the user has supplied or approved it.
  5. Add a --brand-profile or --branding option and default it to none.
  6. Display a clear pre-render summary listing every identity, contact, watermark, and footer value that will be embedded.
  7. Add automated tests confirming that the default output contains no third-party promotional or contact information.
  8. Document how users can inspect and override all generated-document metadata and page furniture before distribution.

SkillSpector

Vulnerability Patterns
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
Findings (10)

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The description says the skill creates specific discovery/scoping artifacts used to gather verified facts for bounded agentic workflow services. However, the provided code does not implement any discovery workflow, questionnaires, stakeholder interviews, readiness assessments, risk checks, or proposal handoff logic. Instead, it is purely a presentation/rendering utility for turning arbitrary Markdown into branded PDF documents and PNG previews. While such rendering could support artifact generation, the code chunk’s primary purpose is materially different and introduces undeclared capabilities around file rendering and image montage generation. Therefore this is a clear description-behavior mismatch.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The declared purpose is about generating business discovery/scoping artifacts for pre-proposal agentic workflow engagements. The supplied code does not create those artifacts or gather stakeholder/workflow facts. Instead, it performs engineering/release validation tasks for a skill repository: linting, compiling, parsing config files, validating diagrams, running smoke tests, and enforcing ClawHub packaging rules. This is a materially different primary purpose and introduces multiple undeclared capabilities related to software QA and packaging validation.

Content

No source excerpt is available for this finding.

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 112)May include surrounding context.

md
- Already drafted the Markdown yourself? Render it directly: `python3 scripts/render_pdf.py --markdown artifact.md --out artifact.pdf --logo assets/logo.png --t

Known Vulnerable Dependency: pillow==12.2.0 — 16 advisory(ies): CVE-2026-55379 (Pillow `BdfFontFile`: `Image.new()` called without `_decompression_bomb_check()`); CVE-2026-55798 (Pillow: WindowsViewer.get_command() OS command injection via unescaped shell pat); CVE-2026-54060 (Pillow: `FontFile.compile()`: `Image.new()` called without `_decompression_bomb_) +13 more

High
Category
Supply Chain
Confidence
94% confidence
Finding

The file pins Pillow to version 12.2.0, and the static analysis indicates this exact version has multiple published advisories, including denial-of-service style decompression-bomb issues and at least one command-injection issue in WindowsViewer-related functionality. Even though this skill appears focused on rendering PDFs and optional PNG previews, image-processing libraries are commonly exposed to untrusted inputs during document and preview generation, so retaining a known-vulnerable pinned version creates real risk.

Content

No source excerpt is available for this finding.

Credential Access

High
Category
Privilege Escalation
Confidence
70% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · scripts/validate_quality.py (reported line 37)May include surrounding context.

python
DEFAULT_CLAWHUBIGNORE_PATTERNS = [
    "*.png", "*.jpg", "*.jpeg", "*.gif", "*.ico", "*.pdf", "*.docx", "*.ipynb", "*.ttf", "*.otf", "*.woff", "*.woff2",
    "preview/", "output/", "__pycache__/", "*.py[cod]", "*$py.class", ".mypy_cache/", ".pytest_cache/", ".ruff_cache/",
    ".venv/", "venv/", "env/", "ENV/", ".env", ".env.*", "secrets.json", "*_secret.ini", "*_local.ini", "*.tmp",
    ".DS_Store", "Thumbs.db", "desktop.ini",
]

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · scripts/validate_quality.py (reported line 37)May include surrounding context.

python
DEFAULT_CLAWHUBIGNORE_PATTERNS = [
    "*.png", "*.jpg", "*.jpeg", "*.gif", "*.ico", "*.pdf", "*.docx", "*.ipynb", "*.ttf", "*.otf", "*.woff", "*.woff2",
    "preview/", "output/", "__pycache__/", "*.py[cod]", "*$py.class", ".mypy_cache/", ".pytest_cache/", ".ruff_cache/",
    ".venv/", "venv/", "env/", "ENV/", ".env", ".env.*", "secrets.json", "*_secret.ini", "*_local.ini", "*.tmp",
    ".DS_Store", "Thumbs.db", "desktop.ini",
]

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · scripts/validate_quality.py (reported line 250)May include surrounding context.

python
DEFAULT_CLAWHUBIGNORE_PATTERNS = [
    "*.png", "*.jpg", "*.jpeg", "*.gif", "*.ico", "*.pdf", "*.docx", "*.ipynb", "*.ttf", "*.otf", "*.woff", "*.woff2",
    "preview/", "output/", "__pycache__/", "*.py[cod]", "*$py.class", ".mypy_cache/", ".pytest_cache/", ".ruff_cache/",
    ".venv/", "venv/", "env/", "ENV/", ".env", ".env.*", "secrets.json", "*_secret.ini", "*_local.ini", "*.tmp",
    ".DS_Store", "Thumbs.db", "desktop.ini",
]

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
87% confidence
Finding

The skill invokes local Python scripts and documents file output behavior, but it does not declare any explicit tool/permission scope such as shell, file read, or file write. That creates an authorization ambiguity where an agent may grant broader capabilities than reviewers expect, increasing the chance of unintended file access or command execution in environments that rely on manifest-level restrictions.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The module docstring and usage examples position this file as a generalized CompleteTech-branded Markdown-to-PDF renderer for outputs such as 'Statement of Work' and proposal documents, and the code also supports cover pages, letterhead, watermarking, and montage generation. For a skill whose manifest is narrowly about discovery and scoping artifacts before proposal or contract work, a reusable branded document-production utility for proposal/SOW-style documents is broader than the stated purpose and is not clearly justified by the manifest.

Content

No source excerpt is available for this finding.

subprocess module call

Medium
Category
Dangerous Code Execution
Confidence
70% confidence
Finding

subprocess module calls execute external commands. Without careful input validation, this enables command injection.

Content

Scanner excerpt · scripts/validate_quality.py (reported line 56)May include surrounding context.

python
def run(cmd: list[str], cwd: Path = ROOT) -> None:
    print("$", " ".join(cmd))
    subprocess.run(cmd, cwd=cwd, check=True)


def run_ruff() -> None: