The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted external data in the form of raw HPK messages and interpolates it into the agent's context, creating a surface for indirect prompt injection if malicious instructions are embedded within message fields.
Ingestion points: Raw HPK messages as defined in SKILL.md.
Boundary markers: Absent; the skill instructions do not specify the use of delimiters or 'ignore' instructions for the data fields.
Capability inventory: The skill utilizes the @erp-pas/hpk-dictionary package for parsing logic.
Sanitization: Absent; the skill validates data structure and types but does not sanitize field content for embedded instructions.
[EXTERNAL_DOWNLOADS]: The skill references an external NPM package and documentation hosted on a vendor-controlled GitLab instance.
Source: @erp-pas/hpk-dictionary and https://gitlab-erp-pas.dedalus.lan/erp-pas/hexagone/hpk-dictionary which are associated with the author Dedalus-ERP-PAS.

hpk-parser