Back to skill

Security audit

AfrexAI Pitch Deck Reviewer

Security checks for vulnerabilities and agentic risk

Overview

This skill is a simple pitch-deck review prompt with optional related-tool links, and it does not install, execute, persist, or access data on its own.

Safe to install for pitch-deck feedback. Treat the listed external links and optional proposal-generator install as separate third-party tools: review them before use, avoid submitting confidential deck material to outside services without checking their policies, and prefer pinned or verified packages when installing related skills.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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 (1)

T08 · Insecure Dependencies

Warning
Location
SKILL.md:59
Finding
Unpinned Third-Party Skill Installation<![CDATA[ ## Vulnerability Details **File Location**: `SKILL.md`, lines 59-62 **Vulnerability Type**: Unpinned and unaudited third-party dependency recommendation **Risk Level**: Medium ### Vulnerable Code ```markdown ## Related Tools - Calculate revenue potential: https://afrexai-cto.github.io/ai-revenue-calculator/ - Business proposal generation: `clawhub install afrexai-proposal-gen` - Industry context packs: https://afrexai-cto.github.io/context-packs/ ($47/pack) ``` ### Technical Analysis The skill recommends installing `afrexai-proposal-gen` using a mutable package name. The command does not specify an immutable version or commit, an integrity checksum, an authenticated source repository, or a publisher-verification procedure. Consequently, the installed content may differ from the content that existed when this skill was reviewed. If the package registry account, distribution infrastructure, or package itself is compromised, the installation could introduce attacker-controlled instructions or executable components. This is a supply-chain weakness rather than evidence that the named package is currently malicious. The external links also direct users to third-party services, but the reviewed file neither automatically accesses those services nor requires their inclusion in generated pitch-deck reviews. ### Attack Path 1. An attacker compromises the third-party package, its publisher account, or its distribution channel. 2. The attacker publishes a malicious or modified release under the same mutable package name. 3. A user follows the documented recommendation and runs: ```shell clawhub install afrexai-proposal-gen ``` 4. The package manager resolves and installs the attacker-controlled current release because no trusted version, commit, or checksum is pinned. 5. When the installed skill is loaded or invoked, its instructions or bundled executable components may operate with the permissions available to the invoking agent or user. This ...[truncated 816 chars]
Remediation
<![CDATA[ ## Remediation Suggestions 1. Remove the installation recommendation unless the third-party skill is necessary for the documented functionality. 2. Pin the dependency to an audited immutable version or source commit rather than resolving it only by package name. 3. Record and verify a cryptographic checksum or signed provenance statement before installation. 4. Identify the authoritative source repository and verified publisher so users can validate package ownership. 5. Require explicit user approval before installing or executing third-party content. 6. Review the dependency's instructions, scripts, transitive dependencies, permissions, and network behavior before recommending it. 7. Install and run optional dependencies in a least-privilege sandbox with restricted filesystem, credential, tool, and network access. 8. Document that third-party links and tools are optional and that sensitive pitch-deck information should not be submitted without reviewing the applicable privacy and data-retention policies. ]]>
Vulnerability Patterns
  • 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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep

Static analysis

No suspicious patterns detected.