Skip to content

fix: soul.md silently truncated to 2000 chars in agent system prompt #672

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
yaojin3616 merged 1 commit into from
Jun 11, 2026
+7 −1
Merged

fix: soul.md silently truncated to 2000 chars in agent system prompt #672

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 7 additions & 1 deletion backend/app/services/agent_context.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -250,7 +250,13 @@ async def build_agent_context(agent_id: uuid.UUID, agent_name: str, role_descrip
- Database → relationship network (human + agent)
"""
# --- Soul ---
soul = await _read_file_safe(normalize_storage_key(f"{agent_id}/soul.md"), 2000)
# Soul is the agent's full author-curated identity; detailed souls (e.g.
# bundle agents) run 4-12k chars. A tight cap silently drops every tail
# section — rules, boundaries, facts — and the agent then confidently
# denies things its soul plainly states, with no log of the truncation.
# Memory and relationships below keep small caps because they grow
# unbounded at runtime; the soul does not (only seeded/explicitly edited).
soul = await _read_file_safe(normalize_storage_key(f"{agent_id}/soul.md"), 30000)
# Strip markdown heading if present
if soul.startswith("# "):
soul = "\n".join(soul.split("\n")[1:]).strip()
Expand Down