The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest untrusted data in the form of user-supplied requirements and constraints to generate architecture recommendations and Architecture Decision Records (ADRs).
Ingestion points: Requirements, constraints, and context gathered from the user in Step 1 of the workflow (SKILL.md).
Boundary markers: Absent; the skill does not specify delimiters or instructions to ignore embedded commands within the user-supplied data.
Capability inventory: The skill instructs the agent to write files to the local file system, specifically ADRs in docs/decisions/ and diagrams in docs/architecture/ (SKILL.md).
Sanitization: Absent; there is no mention of filtering or escaping user input before it is used to generate output files.
[PROMPT_INJECTION]: No evidence of direct prompt injection, jailbreak attempts, or instructions to override agent safety guardrails was found in the skill instructions.
[DATA_EXFILTRATION]: No commands for accessing sensitive system files (e.g., SSH keys, credentials) or unauthorized network exfiltration patterns were detected. The included code snippets are benign examples.
[REMOTE_CODE_EXECUTION]: No remote script downloads or dynamic execution of external code were found. The skill relies on local reference files for its logic.
[OBFUSCATION]: Several reference files (e.g., references/architecture-patterns.md, references/performance-optimization.md) contain 150 sections of highly repetitive technical boilerplate. While unusual and potentially deceptive in nature, no hidden malicious commands or obfuscated URLs were detected within the repeated text.

solution-architecture