Back to skill

Security audit

Blog Figure Svg

Security checks for vulnerabilities and agentic risk

Overview

This skill is a coherent blog-figure generator with some hardening cautions, but no evidence of hidden, destructive, or deceptive behavior.

Install only if you want an editorial SVG/PNG generation workflow. Use a virtual environment or managed rasterizer, prefer pinned dependencies, and keep accent/color inputs to known hex values from the documented palette.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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
Findings (2)

T09 · Insecure Skill Coding Practices

Warning
Location
SKILL.md:347
Finding

Unvalidated Accent Value Permits SVG Attribute Injection

Content
View full analysis
{m.group(1)}' ) else: hero = ( f'' f'' f'' f'' f'' ) svg = f''' {html.escape(headline)} Feature card for blog post: {html.escape(headline)}. Pill label: {html.escape(pill)}.
Remediation
View remediation

T08 · Insecure Dependencies

Note
Location
SKILL.md:39
Finding

Unpinned CairoSVG Installation Creates Supply-Chain Exposure

Content
View full analysis
Remediation
View remediation
" ``` 2. Maintain a lock file containing exact versions for CairoSVG and all transitive dependencies. 3. Require package hashes, for example through a hash-locked requirements file: ```bash python3 -m pip install --require-hashes -r requirements.txt ``` 4. Install dependencies inside a dedicated virtual environment or isolated build container rather than the user's global Python environment. 5. Use a trusted package index and review dependency changes before updating the lock file. 6. Where practical, prefer a preinstalled, organization-managed rasterizer rather than installing packages dynamically during Skill execution. 7. Run rasterization with least privilege and without access to unrelated credentials or sensitive project directories. ]]>
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The trigger list ends with "or any request to produce a chart/diagram for editorial use," which is broad and not tightly constrained to explicit invocation phrases. This can overlap with common content-editing requests and may cause unintended activation because the boundaries of what counts as editorial use are not clearly limited.

Content

No source excerpt is available for this finding.