hacktricks-api

HackTricks AI API (cached, rate-limited)

Escalation layer over the local hacktricks skill. Talks to the HackTricks AI RAG/chat API. Always try the local hacktricks search first — this endpoint has a weekly RAG quota.

Setup (one-time)

Key lives in env HACKTRICKS_API_KEY (already set). Never pass it on the command line or print it. Optional: HACKTRICKS_CACHE (default D:\knowledge\hacktricks\.cache), HACKTRICKS_USAGE_FLOOR (default 2).

Usage

python scripts/htk_query.py usage                      # check remaining weekly quota (free, no spend)
python scripts/htk_query.py rag "AWS IAM privilege escalation" --max-results 5
python scripts/htk_query.py chat "explain SeImpersonate abuse"     # conversational (shares the weekly cap)

• rag returns the top results[] (content + score + filename) plus a cached flag.
• Identical rag queries are served from a content-hash cache (0 HTTP, 0 quota).
• Before any spend the client checks GET /rag/usage and refuses if remaining <= floor — it tells you to use the local hacktricks skill instead.
• The API key is read only from the env and is redacted from all output/errors (including HTTP error bodies).

Notes

• /chat shares the same weekly cap and is not cached (conversational); use rag for repeatable lookups.
• Content is HackTricks (CC BY-NC-SA) — personal reference only.