The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: Instructions in SKILL.md guide the user to clone a third-party repository (https://github.com/codingo/NoSQLMap.git) and run its installation script (python setup.py install). This pattern facilitates the execution of arbitrary code from a source that is not recognized as a trusted organization, posing a significant security risk to the user's environment.
[EXTERNAL_DOWNLOADS]: The skill relies on external tools and documentation from book.hacktricks.xyz, which has been flagged as malicious by automated security scanners (URL:EshopScam). Furthermore, Step 5 in SKILL.md recommends pip install nosqli without version pinning, which exposes the user to potential supply chain attacks.
[INDIRECT_PROMPT_INJECTION]: The automation scripts scripts/agent.py and scripts/process.py ingest raw HTTP response data from external targets to determine if an injection was successful (e.g., checking for strings like 'dashboard' or 'welcome'). This creates a vulnerability where a malicious target server could return crafted content designed to influence the agent's internal logic or manipulate the resulting assessment report.
Ingestion points: requests.get and requests.post calls in agent.py and process.py that read resp.text.
Boundary markers: None identified; response data is directly evaluated against string literals.
Capability inventory: Network operations (requests), filesystem writes (nosql_injection_report.md).
Sanitization: None; external data is processed as raw text without escaping or validation.
[METADATA_POISONING]: There is a discrepancy between the author listed in the YAML frontmatter ('mahipal') and the copyright holder in the LICENSE file ('mukul975'), which may indicate inconsistent or deceptive metadata management.

exploiting-nosql-injection-vulnerabilities