fix(dream): ground memory audit records in the real git diff (#4673)

* fix(dream): ground commit messages and cursor advance in the real git diff

Dream consolidation could emit a /dream-log audit record that did not match
the actual file changes: build_dream_commit_message appended the LLM's
unverified resp.content, dream_run_completed only checked the stop reason, and
file contents were deliberately omitted from the prompt. The combination let a
single-turn self-report become the durable audit record.

- gitstore: add summarize_working_tree() — a structured, machine-derived
  summary (per-file +N/-M, totals, capped unified diff) of working-tree
  changes vs HEAD. Pure filesystem/git ground truth, never LLM narrative.
- memory: build_dream_commit_message now takes the diff body instead of resp;
  dream_content_diff() exposes the real delta over SOUL/USER/MEMORY.md only
  (excludes .dream_cursor so cursor writes aren't mistaken for edits);
  build_dream_prompt embeds current file contents so the model edits reality,
  not a stale mental model.
- builtin/cli: both Dream paths now compute the diff, gate cursor advance on
  a non-empty delta (no-op runs no longer swallow history), and commit with
  the diff-grounded message. Non-git workspaces fall back to the completion
  check.
- dream.md: document that contents are embedded, and add a chain-of-
  verification guardrail so the model's summary cannot claim unmade edits.

A regression test proves a lying resp.content never reaches the audit log
while the real diff does.

* fix(dream): mark non-UTF-8 memory files as binary in diff summary

Address review feedback (Q1 on PR #4673): summarize_working_tree read
working-tree files with errors="replace", which would emit U+FFFD
replacement chars into the audit record if a memory file ever held
invalid UTF-8 — misrepresenting the diff it is meant to make truthful.

Switch to errors="strict" and catch UnicodeDecodeError: a non-UTF-8
(or binary/corrupt) file is now recorded as "{path}: binary or
non-UTF-8 file changed" and omitted from the unified diff, so the
audit record stays honest. An empty diff block is also suppressed when
all changes are binary.

Adds a defensive regression test asserting no replacement char leaks.
This commit is contained in:
Kenneth Zhao
2026-07-06 12:12:55 +08:00
committed by GitHub
parent 70505bd1fc
commit f0c989ba2d
8 changed files with 435 additions and 40 deletions
+11 -2
View File
@@ -359,6 +359,7 @@ async def cmd_dream(ctx: CommandContext) -> OutboundMessage:
store = loop.context.memory
content = ""
resp = None
diff_body = ""
t0 = time.monotonic()
try:
result = store.build_dream_prompt()
@@ -379,9 +380,17 @@ async def cmd_dream(ctx: CommandContext) -> OutboundMessage:
on_progress=_silent,
)
elapsed = time.monotonic() - t0
if MemoryStore.dream_run_completed(resp):
# Ground truth: the real file delta, not the LLM's self-report.
diff_body = store.dream_content_diff()
productive = bool(diff_body) or (
not store.git.is_initialized()
and MemoryStore.dream_run_completed(resp)
)
if productive:
store.set_last_dream_cursor(last_cursor)
content = f"Dream completed in {elapsed:.1f}s."
elif MemoryStore.dream_run_completed(resp):
content = f"Dream completed in {elapsed:.1f}s; no memory changes."
else:
content = (
f"Dream did not complete after {elapsed:.1f}s; "
@@ -399,7 +408,7 @@ async def cmd_dream(ctx: CommandContext) -> OutboundMessage:
timezone_name=getattr(loop.context, "timezone", None),
)
if store.git.is_initialized():
commit_msg = build_dream_commit_message("dream: manual run", resp)
commit_msg = build_dream_commit_message("dream: manual run", diff_body)
sha = store.git.auto_commit(commit_msg)
if sha:
content += f" (commit {sha})"