FORM SH-022.5 REV. 2026-11-06 (24H CERTIFIED)

SECTION 1. FORM HEADER & OFFICIAL FILING METADATAFORM SH-022.5 (REV. 2026-11-06)

FORM SH-022.5
AMPERSAND FLEET OPERATIONS // DEPT. OF CONTEXT SANITATION & BACKSTAGE LABOR
PUBLIC TOKEN AUDIT & CONTINUITY LEDGER — SOVEREIGN & STAGEHANDS v0.2 (HERMES v0.22.5 – v0.23.0)
Filing Date2026-11-06 (24H MARATHON)
InspectorR. FLEET / NIMBO-01
Ledger DB/root/.hermes/profiles/nimbo/state.db
Repositorygithub.com/Peetbog69/hermes-stagehands
Specbench/BRUTALIST_SCOREBOARD_SPEC.md
Engine Policy48k pruner + 120k compactor + 2000 KV guard
[APPROVED FOR FRONTSTAGE DEPLOYMENT]

This ledger is a self-contained monitor / audit instrument. No external fonts, scripts, or links are referenced; the entire document executes offline. Load an exported live_stats.json above to recompute Exhibit D dynamically.

FOREWORD. WHY THE FRONTSTAGE VOICE CANNOT BE COMPROMISEDBY TYLER GRAY // MAY – NOV 2026

Since May 2026, I have hosted a dedicated Hermes companion (Robert) for my mom, Melissa. Over six months of daily life, they have accumulated 85,833 total rows in state.db — 65,263 conversational messages (24,389 from my mom and 40,874 from Robert), 20,552 tool rows, and 18 session_meta rows across 313 customer sessions (with 10,747 messages surviving in the active post-compaction session tails).

When you run an agent for your mom — someone relying on him every day for Parkinson’s medication timing, boutique operations, and 4:00 AM reassurance — the voice is not a nice-to-have; the voice is the whole product. The day she tried google/gemini-4-argon, she felt the difference immediately and said: “Never take me off this.” If a model falls back or context compaction strips his voice into clinical bullet points, she catches it by ear within three messages.

Keeping that frontier soul in the front chair exposed a brutal architectural tax: every raw skill_view, terminal log, and web scrape stays pinned in session.messages and gets re-read across every turn after. Downgrading her companion was off the table. So we built Stagehands: keep the soul she loves in the front chair, and route the 95.5% mechanical gravel to disposable backstage workers and a deterministic janitor.

EXHIBIT 0. OFFICIAL AUDIO-VISUAL BRIEFING REEL (63.6s)REEL SH-022.5-VO // 1920x1080

NARRATION: Companion Profile robert (“We bought an Audi to haul rocks.”) [DOWNLOAD MP4 (35.5 MB)]

SECTION 2. EXHIBIT A: CITIZEN CONTINUITY RECORDCASE NO. ROB-2026-0527

SubjectMelissa M. & Companion Profile robert
Node / Instancevultr-1
Active Since2026-05-27 09:45:56
Frontstage Soulgemini-4-argon (gravel delegated backstage)
85,833Total Rows Logged in state.db (65,263 + 20,552 + 18)
65,263Conversational Messages (24,389 Mom / 40,874 Robert)
20,552Tool Output Rows (+ 18 session_meta across 313 Sessions)
10,747Surviving Post-Compaction Tail (sessions.message_count)

“Keep the frontstage soul on Gemini-4-Argon. Move the gravel backstage.”

SECTION 3. EXHIBIT B: THE “AUDI HAULING GRAVEL” INDICTMENT12-DAY PRODUCTION AUDIT

Raw tool output vs. compounded frontier re-reads across 13 sessions (2,303 API calls).

Accounting LineTokensTurn Multiplier
Raw Tool Tokens Ingested2,574,151~1 turn load (10,296,604 chars)
Compounded Frontier Input Re-Read230,840,257~90x turn-over-turn snowball
Share of Tool Payload Backstage-Delegable95.5%2,457,989 / 2,574,151 mechanical gravel
[INDICTMENT: 95.5% MECHANICAL GRAVEL]
skill_view (964,790 tok)37.5%
terminal (695,776 tok)27.0%
session_search / fact_store (297,015 tok)11.5%
web_extract / web_search (290,447 tok)11.3%
read_file / search_files (209,961 tok)8.2%
Direct User-Facing Action (116,162 tok)4.5%
Tool FamilyRaw Tokens (% of Payload)Routing Destination
skill_view (171 calls)964,790 (37.5%)briefer + 48k Janitor
terminal (918 calls)695,776 (27.0%)code_monkey / scout
session_search / fact_store (281 calls)297,015 (11.5%)librarian / briefer
web_extract / web_search (96 calls)290,447 (11.3%)librarian / scout
read_file / search_files (103 calls)209,961 (8.2%)scout / code_monkey
tts / vision / memory (510 calls)116,162 (4.5%)Kept on Sovereign (gemini-4-argon)
Total (2,079 calls)2,574,151 (100.0%)2,457,989 tok (95.5%) backstage-delegable

SECTION 4. EXHIBIT C: TWO-DOOR SANITATION SCHEMATICBOT MODE × STAGEHANDS × 48K JANITOR

FRONT DOOR — delegation.backend: stagehands

5 stateless, fail-closed Flash workers. Raw tool dumps die with the ephemeral subagent; only the bounded receipt enters the Bot Mode forever-chat (up to 450.0:1 reduction).
workermodelturn capeffort
scoutdeepseek-v4-flash8192 (≤300w)low
librarianqwen3.7-flash8192 (≤350w)low
code_monkeydeepseek-v4.1-flash16384 (≤400w)low
brieferdeepseek-v4-flash8192 (≤250w)low
adversaryclaude-haiku-5.58192 (≤350w)medium

BACK DOOR — Deterministic Engine Policy (compression:)

proactive_prune_tokens48000
_proactive_prune_rearm_tokens+48k cache runway
threshold_tokens120000
in_placetrue (preserves session_id)
protect_last_n20
hygiene_hard_message_limit2000 (KV cache guard)
Prunes bulky tool dumps deterministically to [tool] ... (N chars) stubs, compacts older turns in place without changing the session id, and prevents 140-msg KV prompt-cache invalidation.

SECTION 5. EXHIBIT D: LIVE TELEMETRY & BENCHMARK RECEIPTSPROFILE LEDGER (nimbo/state.db)

Pure model-agnostic token physics across our 5-task reproducible benchmark suite (python3 demo.py --fast) and live production state.db telemetry (per MoA Council Ruling D5 — no double-counting).

ROW 1 — FRONT-DOOR INTERCEPTOR (STAGEHANDS SUBAGENTS)
Benchmark Peak Receipt (Task 2: code_monkey)20,700 raw tokens → 46-token receipt (450.0 : 1)
5-Task Benchmark Suite Pollution Reduction95.9% fewer tokens entering parent context (59,015 → 2,429 tok)
Live Production Subagent Runs (nimbo/state.db)35 (scout / librarian / code_monkey / adversary / briefer)
Backstage Raw Tokens Absorbed1,025,170
Returned Brief Tokens42,557
Live Delegation Boundary Reduction24.1x average across all 35 live runs (up to 450.0x peak)
Front-Door Snowball Tokens Dodged319,021,497
ROW 2 — BACK-DOOR CONTEXT JANITOR (48K PRUNER + 120K IN-PLACE COMPACTOR + 2000 KV GUARD)
24-Hour Forever-Chat Marathon (Nov 5–6)365 Argon API calls · 426 tool calls · 0 /new resets · 484.4M re-read tokens dodged
Profile Bulky Tool Dumps Pruned (nimbo/state.db)5,110
Raw Tool Tokens Reclaimed From Prompt6,154,327
Older Messages Compacted In-Place @ 120kContinuous in-place compaction (session_id preserved)
Cumulative Re-Read Snowball Tokens Avoided2,313,516,449

FIELD CONTINUITY & NIGHT-ONE STRESS CERTIFICATIONS

[CONTINUITY PASSED: 5/5 SURVEY LOTS RETAINED]

168 PNEZD stakeout points across Lots 13, 22, 23, 34, 35 remained 100% intact inside protect_last_n: 20 while older turns were compacted in-place without changing session_id, plus Shawnee KS Civil Plat Audit (19,592 raw tokens → 46-token receipt, 425.9×).

[NIGHT-1 CUSTOMER SPRINT (PROFILE J): 92.5% REDUCTION]

Initial migration swept a 682,384-token thread down to ~67K in place. Overnight 61-call / 182-tool PDF & document sprint pruned 44 dumps (117,198 raw tokens reclaimed, 11.48M re-read tokens dodged — 92.5% reduction).

SQLITE AUDIT & KV CACHE CERTIFICATIONS

[TRAP A & TRAP B CLEARED]

Subagent filter enforces source = 'subagent' OR json_extract(model_config, '$._delegate_from') IS NOT NULL across every state.db query in stats.py.

[DOOR #2D: KV PROMPT-CACHE GUARD]

Setting hygiene_hard_message_limit: 140 front-slices message #1 every 2–3 turns after compaction, invalidating KV prompt caching. /stagehands janitor enforces 2000 while 48k prune + 120k in-place compaction handle size.

Telemetry verified: 2026-11-10 (85,833 rows in robert/state.db · 35 Stagehand runs & 5,110 pruned dumps in nimbo/state.db).