FOUND BY E2E, NOT BY REASONING. The previous commit's design note asserted the provenance
receipt "never reaches the user: it is appended to the history copy, not the reply." That
was FALSE, and only running the thing showed it. On the first live run against the built
DMG brain, two agentic turns out of two came back with
"Your favourite colour is chartreuse and your project is called Perihelion.
[[RECEIPT - recorded by the soul, not written by the model: no tools ran on this turn.]]"
— the receipt in the user-visible reply.
MECHANISM: the receipt is stored inside the assistant turn, and the agentic path replays
history VERBATIM as Anthropic message objects. So the model sees its own previous answers
ending in [[RECEIPT ...]] and does the obvious thing — it imitates the format and signs the
next answer the same way. The plain path did NOT leak, which is the tell: there, history is
rendered into the SYSTEM prompt as labelled lines rather than replayed as assistant turns,
and a model imitates its own turns far more readily than a transcript.
FIX, two layers, because one of them is not a guarantee:
- receipt_rule() names the marker in both system prompts (plain and agentic): these lines
are written by the system, read them as evidence, never write one. Reduces occurrence.
- receipt_strip() truncates any [[RECEIPT ...]] out of model output before it becomes the
reply — plain path in layered_generate, agentic path on final_text in agentic_loop.
Deterministic. A guard that depends on the model choosing to obey is exactly the class of
thing round 8 exists to stop shipping, so the instruction is the optimisation and the
strip is the guarantee.
Placed ABOVE agentic_loop's empty-check on purpose: a turn whose entire output was an
imitated receipt has produced no answer, and must be reported as no answer.
The receipt stays in HISTORY, which is the whole point and is proven to work: asked "What
source did you use for that?" one turn after a live web_search, this brain answered
"I used Weather Underground (https://www.wunderground.com/weather/is/reykjav%C3%ADk) for the
current temperature in Reykjavik" — a real source, no apology. That is the false confession
dead, and it is dead BECAUSE the model can read the receipt.
BUILT: 887,112 bytes, sha256 54a2eff84d4fa44f8d2db6781dcf225df4b5075a8ec5f1058018bd40cc1af10b
BUG-PLAINCHAT-1 miscompile guard: zero sites.
Refs neuron#109
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
DESIGN FIT: three of round 7's five defects share ONE root — the conversation-history
layer persists only {role, content}, discarding tool provenance, session scoping, and the
distinction between a real user turn and an internal utility call. Fixes A and B RESTORE
Will's design rather than extend it: his agentic path already scopes history per session,
the plain path never got it, and his own source carries the TODO admitting the resulting
race (chat.el, handle_chat: "process-global key; concurrent /api/chat requests without
session_id race on this read-append-write"). Fix C repairs one join Will wrote that was
correct for a year and one we added last week. E1/E2 are ours.
FIX A — tool provenance in history (kills the FALSE CONFESSION)
Root cause, EXECUTED-verified: handle_chat_agentic recorded turns via hist_append, which
emits {"role","content"} only. server_tool_use blocks, web_search_tool_result blocks and
every citation were discarded, then replayed as text. On the next turn the model saw a
data-rich answer with zero evidence a search had happened, and its own permanent rule
("never describe a search you did not perform") left one conclusion available: that it
had fabricated the data. It apologised for a search it HAD run — four independent lines
of evidence confirm the search was real. The defect is not the model's honesty. It is
that we deleted the evidence and then asked it to account for itself.
Change: agentic_loop accumulates the source URLs it already walks past (citations and
web_search_tool_result content) and returns them as "sources"; handle_chat_agentic folds
tools_used + sources into a receipt line stored WITH the assistant turn. Receipts are
unconditional — a negative receipt ("no tools ran") is the other half of the guarantee,
because "no evidence of a tool" and "evidence of no tool" were previously identical in
the transcript. conv_history_block splits the receipt off before snipping so a long
answer cannot truncate away the evidence. The user never sees it: it is appended to the
history copy, not the reply.
FIX B — one history key for both paths (kills the BLANK STARE)
Root cause, EXECUTED-verified: the agentic path keyed history on session_hist_<id>; the
plain path was hard-wired to the process-global conv_history and never read session_id.
One conversation, two buckets. Proven in the guest engram: the scoped node held exactly
two turns starting at "Try again" while the earlier exchanges sat unscoped.
Change: conv_hist_key/conv_hist_label are now the single definition, used by BOTH paths;
session_id is threaded route -> layered_cycle -> layered_generate / conv_history_record.
The 2-line fallback (plain path reads the agentic key) was REJECTED: it keeps the
process-global bucket as a live write target, which is the bleed the TODO describes.
Also found and closed while threading: layered_cycle read session_id from the state key
"current_session_id", which is read here and WRITTEN NOWHERE in the entire source. It
was unconditionally "", so TODO(reliability #4) — per-session steward continuity — was
dead code that could never fire. It fires now.
LAZY SESSION, decided explicitly: we create the session EAGERLY at the door (app half,
ui#223) rather than migrating orphaned turns. Migration would copy the CONTENTS of a
process-global bucket, possibly another conversation's, into a named session — the bleed,
performed deliberately. Eager creation makes the situation impossible instead. Migration
is deliberately not implemented and must not be added without solving provenance first.
FIX C — the two text-join seams ("to.Good", byte-verified 0x77 0x2e 0x47)
Two bare `+` joins, written a year apart, had drifted into two answers to one question:
within-response block joins (Will's, 2026-05-03, latent until server-side web_search
began interleaving non-text blocks) and across-round joins (ours, 62af564).
Change: one named rule, text_join_sep, at both sites. NOT a blanket separator — a cited
answer splits MID-SENTENCE ("The current temperature is " + "86°F" + ", with "), so a
blanket separator shatters every sourced sentence. The rule takes the one bit that
distinguishes the cases: whether a NON-TEXT block intervened. Hoisting it also makes the
fix verifiable in the shipped binary, which an inline `+` is not.
FIX E1 — utility generations stay out of the transcript
Title generation ("Write a 3-6 word title...") and insight passes ran down the same plain
door as a real message and were recorded as if the user had typed them; the same calls are
the "model":"unknown" rows in usage.jsonl. is_utility_request reads an explicit utility
flag from the app, with the __title__/__insight__ id prefixes as a fallback for older
clients. Answered normally, never recorded.
FIX E2 — OPERATOR IDENTITY is scoped to tool-capable turns
The block (env USER/HOME, closing "This is a hard rule") was prepended to EVERY system
prompt including chat mode. On a Tools:Off turn there is no filesystem in reach, so it
governed nothing and merely supplied the loudest fact in the prompt — which is why the
model opened a fresh conversation with "You're test, on your machine at /Users/test".
Hoisted to operator_identity_block() and gated on !chat_mode. Unchanged wherever a file
or command tool can actually be reached.
ALSO: agentic_loop's per-session history persist had a second hand-rolled copy of
conv_history_persist with a different label expression, different salience scores and
different tags for the same node. Since both now derive the label from conv_hist_label and
engram_node_full upserts by label, two score policies were writing one node. Collapsed to
one writer.
BUILD NOTE: dist/elp-c-decls.h is force-included by the documented link recipe and carried
the OLD C arities, so it is updated here. This is the build-support header, NOT the stale
generated dist/soul.c — no dist/*.c was read or edited; all engine changes are .el source.
chat.elh/soul.elh are committed because a first-pass build against the old signatures FAILS
(measured); the other regenerated headers are reverted as unrelated churn.
BUILT: 887,000 bytes, sha256 d632b061ad75269d6adeb52578d030eaf49e895d91289d7f946b19c08450d728
Zero el_str_concat(<int>, str_len(...)) sites (the BUG-PLAINCHAT-1 miscompile guard).
web_search_20250305 and disable_parallel_tool_use both still present — PR #108's web search
and the ADR-0005 stopgap are intact.
Refs neuron#109 (builds on it), neuron#78 (Receipt Contract — the real fix A is a stopgap for)
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The session-init endpoints concatenated unbounded engram activate/scan
results as FULL node objects (content up to ~90KB per node), producing a
~900KB response. After the MCP wrapper re-escapes that into a stringified
text block the client dropped the socket ('connection closed unexpectedly')
on every beginSession call. Cap each list (8-10 activated, 10-20 recent)
and project every node to a light identity plus a bounded, UTF-8-safe
content snippet. Response drops from ~900KB to ~12KB; full content stays
available on demand via recall/fetch/inspectGraph.
- awareness.el: curiosity auto_term was the raw first word of a WM
label with no term-quality scoring — observed seeds included What,
Colon, Prose, Context. Replaced the 7-word genre blocklist whack-a-
mole with a delimited stopword membership test (function words +
document-structure words); topical terms pass untouched. Verified
live: seeds now ReasonEdit, Reasoning-model, Self-review.
- routes.el + awareness.el: idle counter only reset on rare inbox
synthesis-requests, so idle==pulse always (zero information).
handle_request now stamps soul.last_activity_ts on every inbound
HTTP request; heartbeat emits idle_ms = ms since last request
(-1 until first request of a boot).
- neuron-api.el: beginSession concatenated a depth-2 spread plus the
unbounded self-hub neighbor dump — multi-MB response, doubled by
wrapper re-escaping, socket died on every call. Now depth-1 and
the hub dump dropped (identity loading has its own tool). Verified:
beginSession returns instead of closing the socket.
- engram refresh URL now resolves env -> state -> localhost:8742, same
hardening ise_post got after the boot-4 blackout. Previously a corrupted/
empty soul_engram_url state key silently disabled sync forever while
heartbeats kept flowing — WM starves of Knowledge nodes with no outward
sign.
- heartbeat ISE: node_delta, edge_delta (growth vs stall vs flood is now
one field, not cross-ISE forensics), sync_age_ms from a new
soul.last_sync_ok_ts stamp (-1 = never; >> SOUL_REFRESH_MS = refresh
path broken). Verified live: pulse 1 sync_age_ms=-1, sync fired +1.6s,
age counts up between syncs.
- PR #60: inject operator home dir into system prompt (#30)
Adds OPERATOR IDENTITY section so the LLM correctly resolves
'my files/notes/desktop' to the actual running user's $HOME.
Prevents identity confusion between imprint author and operator.
- PR #61: plan-mode endpoint POST /api/chat {mode:'plan'} (#27)
Adds handle_chat_plan — returns {steps:[{id,title,detail}]} JSON.
Wired into all three /api/chat route handlers. Grounds the plan
via engram_compile (same as agentic path) for context awareness.
dist changes:
- soul.c: both PRs compiled in; build_system_prompt updated to
2-param signature (ctx, chat_mode); handle_chat_plan added
- chat.c/routes.c/chat.elh: individual module outputs updated
- elp-c-decls.h: remove stale 1-param build_system_prompt decl,
add handle_chat_plan declaration
- soul.elh.c: new soul header declarations file (from PR #60)
Compile verified: cc -O2 -DHAVE_CURL soul.c el_runtime.c -lcurl
Binary: 805K arm64, smoke test passes (port in use = expected).
neuron.c and routes.c were compiled against the old 1-arg soul interface.
chat.c already uses the 2-arg signature. The Windows cross-compile build
generates elp-c-decls.h from all dist/*.c files, causing a conflicting-types
error when both signatures appear. Recompile these modules against the
current soul API to eliminate the conflict.
Linked against dev runtime with is_knowledge fix that adds Knowledge
node type. Engram goal_bias now gives Knowledge nodes +0.3 boost on
technical queries, consistent with how Belief/DharmaSelf/Safety nodes
are already treated. Same el_runtime source as concurrent foundation/el
commit 16d62bd.
proactive_curiosity() now uses the top working-memory node's first label
word as a 4th activation seed alongside the 4 rotating fixed sets. This
breaks deterministic exploration that was reinforcing the same subgraph
every cycle and creates a self-referencing loop: curiosity radiates from
whatever is most salient right now, mirroring the brain's default-mode-
network resting-state dynamics. str_find_chars on " :([" extracts the
first meaningful word; sp > 3 guards against bracket-prefixed labels.
auto_term field added to curiosity_scan ISE for observability.
Ollama availability is a silent failure mode: when the embedding service
is down, semantic seed injection falls back to lexical-only activation with
no signal in the ISE stream. Add embed_ok field (0/1) to every heartbeat
by probing http://localhost:11434 — makes Ollama health visible without
a separate monitoring path.
state_get() returns "" for unset keys. Both soul.pulse and soul_boot_count
could be empty on first heartbeat cycle, producing invalid JSON like
{"event":"heartbeat","pulse":,"boot":,...}. Add defensive guards:
if str_eq(raw, "") { "0" } else { raw } for both fields.
Tim Lingo
will.anderson