T09 · Insecure Skill Coding Practices
- Location
SKILL.md:227- Finding
Unencrypted HTTP Fallback for arXiv Metadata Retrieval
- Content
View full analysis
Vulnerability Details
File Location:
SKILL.md, line 227
Vulnerability Type: Plaintext external API communication
Risk Level: MediumVulnerable Code
markdown | Semantic Scholar rate limit | Fall back to arXiv API: `http://export.arxiv.org/api/query?id_list=<id>` |Technical Analysis
The skill instructs the agent to use an unencrypted HTTP endpoint when Semantic Scholar is rate-limited. HTTP does not provide transport encryption, server authentication, or response integrity.
An attacker with a suitable network position could observe requested arXiv identifiers or modify the API response in transit. Because the skill subsequently extracts metadata and writes it to
PAPERS.md, a tampered response could become part of the user's persistent research knowledge base. This is an insecure coding and configuration practice rather than remote code execution: the retrieved material is treated as paper metadata, not executable code.Attack Path
- A paper-ingestion request causes the agent to query Semantic Scholar.
- Semantic Scholar returns a rate-limit response.
- Following the documented fallback, the agent requests the arXiv API over HTTP.
- A network-positioned attacker intercepts the plaintext request and supplies altered metadata.
- The agent accepts and analyzes the manipulated response.
- False titles, authors, abstracts, conclusions, or bibliographic details may be persisted in
PAPERS.mdand presented in later sessions.
Impact Assessment
Exploitation requires control of or visibility into the network path, such as a malicious proxy, compromised access point, or upstream network adversary. The attacker does not gain operating-system privileges or direct code-execution capability from this issue.
The affected scope is the confidentiality and integrity of fallback arXiv requests and the persistent research records derived from their responses. Possible consequences includ ...[truncated 199 chars]
- Remediation
View remediation
Remediation Suggestions
- Replace the plaintext endpoint with its HTTPS equivalent:
text https://export.arxiv.org/api/query?id_list=<id> - Prohibit redirects or fallback behavior that downgrades HTTPS requests to HTTP.
- Validate that the final response URL retains the expected
httpsscheme and approved arXiv hostname. - Treat externally retrieved metadata as untrusted input and validate expected response structure, identifiers, and field lengths before persistence.
- Cross-check important metadata, such as title, authors, and arXiv identifier, against an independent HTTPS source before writing it to
PAPERS.md. - If secure retrieval fails, stop and report the failure rather than silently using plaintext transport.
- Replace the plaintext endpoint with its HTTPS equivalent: