The Agent Skills Directory

[OBFUSCATION]: Several reference files (e.g., 'references/architecture-patterns.md', 'references/code-organization.md', 'references/deployment-pipelines.md', 'references/error-handling.md', 'references/performance-optimization.md', 'references/security-best-practices.md', 'references/state-management.md', and 'references/testing-strategies.md') contain a massive amount of repetitive technical filler. Each of these eight files contains exactly 150 sections that repeat and reorder the same small set of paragraphs and code snippets hundreds of times. This 'noise injection' is a characteristic pattern used to overwhelm automated scanners and manual reviewers, potentially hiding malicious payloads or instructions among over 800KB of redundant data.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a significant attack surface for indirect prompt injection due to its ingestion of a high volume of technical instructions from unverified reference files without sanitization or clear boundary markers. 1. Ingestion points: Technical reference files in the 'references/' directory. 2. Boundary markers: Absent in the repetitive deep-dive files. 3. Capability inventory: Subprocess calls via Ansible (eos_config, eos_command) and Nornir (netmiko_send_command) documented in 'SKILL.md'. 4. Sanitization: Absent.
[EXTERNAL_DOWNLOADS]: The skill documents fetching the 'nlnetlabs/routinator' Docker image for RPKI validation in 'references/bgp-anycast.md'. NLnet Labs is a well-known and established organization in the internet infrastructure community, and this download is consistent with the skill's primary purpose.

devops-network-infrastructure