The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill documentation provides examples of passing sensitive credentials as plain-text command-line arguments (e.g., --password changeme). This practice is discouraged as it can lead to credential exposure through shell history, process listings, or environment variables.
[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to how it audits existing security controls.
Ingestion points: The audit_existing_searches function in scripts/agent.py retrieves search metadata and queries directly from the Splunk REST API.
Boundary markers: There are no boundary markers or instructions to ignore embedded commands within the processed data.
Capability inventory: The skill can perform network operations via requests and write files to the local system (report generation and Sigma rule export).
Sanitization: The auditing logic uses simple substring matching on untrusted search content, allowing an attacker to generate false-positive coverage reports by inserting specific MITRE technique IDs into benign search names or comments.
[DATA_EXFILTRATION]: The scripts/agent.py script includes an option to disable TLS certificate verification via the SKIP_TLS_VERIFY environment variable. If enabled by a user, this configuration exposes all communication with the Splunk API, including authentication credentials and security data, to interception or modification via Man-in-the-Middle (MITM) attacks.
[METADATA_POISONING]: There is an inconsistency in the skill's identity metadata. The SKILL.md frontmatter identifies the author as 'mahipal', whereas the LICENSE file attributes the copyright to 'mukul975'. While this may be a clerical error, it represents a lack of consistency in authorship claims.

implementing-siem-correlation-rules-for-apt