The Agent Skills Directory

[COMMAND_EXECUTION]: The shell scripts interpolate user-provided variables like "$TARGET" and "$PROJECT_ID" directly into command strings without sanitization. A malicious input containing shell metacharacters such as ";", "&", or "|" could lead to arbitrary command execution on the host system.
[COMMAND_EXECUTION]: The skill uses the "-k" ("--insecure") flag with "curl" in all network requests, which disables SSL/TLS certificate verification and makes the agent vulnerable to Man-in-the-Middle (MitM) attacks.
[DATA_EXFILTRATION]: The skill is designed to search for and extract sensitive information including CI/CD tokens, API keys, private keys (".pem"), and environment files (".env") from public GitLab repositories.
[EXTERNAL_DOWNLOADS]: The skill fetches metadata and raw file contents from untrusted external GitLab instances. This data is processed and presented to the agent, potentially containing malicious payloads.
[REMOTE_CODE_EXECUTION]: While the script uses "python3 -c" for JSON parsing of "curl" output, the underlying shell command injection vulnerability via the "$TARGET" variable creates a high-risk execution path for remote code.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection. Ingestion points: Raw file contents from GitLab repositories (e.g., ".env", ".gitlab-ci.yml") downloaded via "curl". Boundary markers: None; content is displayed directly to the agent. Capability inventory: The skill can execute shell commands, perform network requests, and write to the local file system. Sanitization: None; content is processed using "grep" and printed without validation or escaping.

gitlab-public-recon