agent-ready-cloudflare
Warn
Audited by Gen Agent Trust Hub on Oct 31, 2026
Risk Level: MEDIUMMETADATA_POISONINGINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [METADATA_POISONING]: The skill's metadata in
SKILL.mdidentifies the author as 'Cloudflare / isitagentready.com', which contradicts the provided uploader context of 'fabricioctelles'. This impersonation of a well-known entity is deceptive and may cause users to misjudge the safety and official status of the skill. - [INDIRECT_PROMPT_INJECTION]: The operational flow described in
SKILL.mdinvolves ingesting raw strings (such as themessagefield) from a remote API and interpolating them directly into prompt templates for the agent. This creates a surface where an attacker-controlled website could return a malicious message that, when presented back to the agent, could override its instructions or behavior. - Ingestion points: API response content from
isitagentready.com/api/scaninterpolated intoSKILL.mdprompt templates. - Boundary markers: No delimiters or safety instructions are used to wrap the
{issue}or{message}placeholders. - Capability inventory: Assembles and presents markdown-based implementation prompts for coding agents.
- Sanitization: No filtering or escaping of the remote API response is specified.
- [COMMAND_EXECUTION]: The skill provides instructions for the agent to construct and execute
curlcommands by substituting aDOMAINplaceholder with user input. The documentation does not require input validation or escaping, which could allow a user (or a malicious site directing the user) to inject additional shell arguments or commands via the domain field. - [EXTERNAL_DOWNLOADS]: The skill performs network operations to
isitagentready.comto fetch site scan data. While this is the core function of the tool and targets a well-known service associated with agentic standards, it represents an external dependency that processes user-supplied URLs through a third-party API.
Audit Metadata