teleo/teleo-infrastructure
8
0
Fork 0
Code Issues Pull requests 1 Projects Releases Packages Wiki Activity Actions 4

epimetheus: sanitize learnings before prompt injection (Ganymede review)

Browse source 
Learnings file content now passes through sanitize_message() before injection
into the Opus prompt. Prevents prompt injection via crafted "corrections."
Rio UUID 5551F5AF confirmed as current Teleo v4 Rio.

Pentagon-Agent: Epimetheus <3D35839A-7722-4740-B93D-51157F7D5E70>
This commit is contained in:
m3taversal 2026-03-21 15:29:46 +00:00
parent 1b4c6f8d72
commit e921eda0a0
1 changed files with 3 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

5
telegram/bot.py
View file

@ -202,9 +202,10 @@ def _git_commit_archive(archive_path, filename: str):
def _load_learnings() -> str:
"""Load Rio's learnings file for prompt injection."""
"""Load Rio's learnings file for prompt injection. Sanitized (Ganymede: prompt injection risk)."""
try:
return Path(LEARNINGS_FILE).read_text()[:3000] # Cap at 3K chars for prompt budget
raw = Path(LEARNINGS_FILE).read_text()[:3000]
return sanitize_message(raw) # Same sanitization as user messages
except Exception:
return ""