@totalreclaw/totalreclaw 3.3.12-rc.4 → 3.3.12-rc.6

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (11) hide show
  1. package/CHANGELOG.md +26 -0
  2. package/SKILL.md +156 -133
  3. package/dist/index.js +45 -5
  4. package/dist/llm-client.js +69 -1
  5. package/dist/tr-cli.js +1 -1
  6. package/import-adapters/types.ts +1 -1
  7. package/index.ts +46 -5
  8. package/llm-client.ts +74 -1
  9. package/package.json +2 -2
  10. package/skill.json +1 -1
  11. package/tr-cli.ts +1 -1
package/CHANGELOG.md CHANGED
@@ -4,6 +4,32 @@ All notable changes to `@totalreclaw/totalreclaw` (the OpenClaw plugin) are docu
4
4
 
5
5
  The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/), and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
6
6
 
7
+ ## [3.3.12-rc.5] — 2026-05-09
8
+
9
+ Final RC for the 3.3.12 stable promote. Behavioral fix: Pedro's Pop-OS Telegram QA (zai/glm-5-turbo) on rc.4 found the agent storing user statements in `MEMORY.md` / `USER.md` via `write` tool calls instead of calling `totalreclaw_remember`. 28 `write` calls observed in one session, 0 TotalReclaw tool calls — facts never reached the chain. Root cause: SKILL.md trigger language was permissive ("call when the user asks") and the agent's default file-write reflex won out. rc.5 makes the memory-storage rule the TOP RULE of SKILL.md with an explicit prohibition on `write`/`edit` against `MEMORY.md`/`USER.md`, an exhaustive trigger-phrase list (preference / identity / decision / commitment / possessive-assertion patterns), and a multi-fact-per-message instruction so blob-style packing does not happen.
10
+
11
+ ### Changed
12
+
13
+ - **SKILL.md TOP RULE rewrite — aggressive triggers, no MEMORY.md fallback.** Memory storage section moved to the very top of SKILL.md (before architecture, vocabulary, install flow). Adds: (1) absolute prohibition on `write`/`edit` against `MEMORY.md` / `USER.md` / `~/.claude/memory/*.md`; (2) trigger-phrase list covering preferences, identity, tools, decisions, commitments, explicit asks, possessive assertions; (3) multi-fact-per-message rule (split into one `totalreclaw_remember` call per atomic fact); (4) full 17-tool reference table with canonical use cases; (5) restart policy reaffirmed (gateway self-restarts via SIGUSR1; agent NEVER prompts user to manually restart); (6) phrase-safety hard rail (no phrase in chat / no phrase as tool input / browser-side only).
14
+ - **Public quickstart guide audit** (`docs/guides/openclaw-setup-quickstart.md`) — already prose-style + clean from rc.6 revert. No edits required.
15
+ - **Long-form setup guide cleanup** (`docs/guides/openclaw-setup.md`) — F-flip default URL corrected (`api.totalreclaw.xyz` for free tier, staging via env override; previous text said the relay was `api-staging.*`); line 1 wording aligned with SKILL.md (`Setting up TotalReclaw — this takes about a minute…`); legacy `qr_png_b64` / `qr_unicode` references replaced with `qr_ascii` (the only QR field in the current pair payload); manual restart fallback (`openclaw gateway restart` / `docker restart`) deemphasized — the plugin self-restarts via SIGUSR1; RC pin examples bumped from `3.3.11-rc.5` to `3.3.12-rc.5`.
16
+ - **Version sync** — package.json, skill.json, SKILL.md frontmatter, tr-cli.ts PLUGIN_VERSION all aligned to 3.3.12-rc.5 via `sync-version.mjs`. `check-version-drift` green.
17
+
18
+ ### Verified
19
+
20
+ - All 38 test suites pass (manifest-shape, config-schema, config, relay-headers, scope-address-visible, llm-profile-reader, llm-client (×2), gateway-url, retype-setscope, tool-gating, onboarding-noninteractive, pair-cli-json, pair-qr, pair-remote-client, qa-bug-report, nonce-serialization, phrase-safety-registry, onnx-download-ux, onboard-pair-only, import-time-smoke, install-staging-cleanup, partial-install-detection, install-reload-idempotency, json-stdout-cleanliness, load-manifest, url-binding, fs-helpers, pair-cli-default-mode, embedding-fallback-tag, staging-banner-gate, restart-auth, inbound-user-tracker, register-command-name, skill-md-hybrid-primary, tr-cli-json-output).
21
+ - `check-scanner` green (0 flags).
22
+ - `check-version-drift` green (3 sites + tr-cli.ts all = 3.3.12-rc.5).
23
+ - E2E on clean OpenClaw 2026.5.7 container — agent given `"Hi, I'm Pedro. I live in Porto. I prefer PostgreSQL over MySQL."` calls `totalreclaw_remember` ≥1 time, with 0 `write`/`edit` calls against `MEMORY.md` / `USER.md`. (See PR description for the captured `toolSummary` JSON.)
24
+
25
+ ### Out of scope (unchanged from rc.4)
26
+
27
+ - Plugin install lifecycle on OpenClaw 2026.5.7 (auto-QA PASS — plugin loads, SIGUSR1 self-restart works, in-process restart, managed-service-friendly).
28
+ - `tr` CLI on-chain memory ops (rc.4 fix verified).
29
+ - Pair flow (browser-side phrase, no leakage).
30
+ - Trajectory poller (auto-extraction backup).
31
+ - F flip (RC + stable both default to prod URL; staging via env override).
32
+
7
33
  ## [3.3.12-rc.2] — 2026-05-08
8
34
 
9
35
  Hot-fix on rc.1's F flip. Pair flow regression: rc.1 set `pairRelayUrl`'s default to `wss://api.totalreclaw.xyz` (production) independently of `serverUrl`. RC users who set `TOTALRECLAW_SERVER_URL=https://api-staging.totalreclaw.xyz` (per the staging-opt-in flow) had pair WS go to **prod**, which pre-dates the pair feature → 404 on WS upgrade → `totalreclaw_pair failed: Unexpected server response: 404`. End-to-end blocker: pair never completed → no credentials → no memories.
package/SKILL.md CHANGED
@@ -1,7 +1,7 @@
1
1
  ---
2
2
  name: totalreclaw
3
- description: "Install + set up TotalReclaw encrypted memory for OpenClaw, then use tr CLI for remember / recall. Trigger on 'install TotalReclaw', 'set up TotalReclaw', 'restore my recovery phrase', or any remember / recall request."
4
- version: 3.3.12-rc.4
3
+ description: "End-to-end encrypted, decentralized memory for OpenClaw. Set up an account once, then call totalreclaw_remember / totalreclaw_recall (or the tr CLI under hybrid-primary) instead of writing to MEMORY.md / USER.md / local files. Trigger on 'install TotalReclaw', 'set up TotalReclaw', 'restore my recovery phrase', any remember / recall request, AND any user statement that contains a preference / fact / decision / commitment about themselves."
4
+ version: 3.3.12-rc.6
5
5
  author: TotalReclaw Team
6
6
  license: MIT
7
7
  homepage: https://totalreclaw.xyz
@@ -24,26 +24,61 @@ metadata:
24
24
 
25
25
  # TotalReclaw — OpenClaw
26
26
 
27
- > **Agent-executable quickstart (5 KB, copy-paste-runnable)**: <https://github.com/p-diogo/totalreclaw/blob/main/docs/guides/openclaw-setup-quickstart.md>
27
+ > **Agent-executable quickstart**: <https://github.com/p-diogo/totalreclaw/blob/main/docs/guides/openclaw-setup-quickstart.md>
28
28
  >
29
- > **Human-readable setup guide (full rationale, troubleshooting)**: <https://github.com/p-diogo/totalreclaw/blob/main/docs/guides/openclaw-setup.md>
30
- >
31
- > If the user pastes either URL: prefer the quickstart for execution. Fetch the full guide only if the user asks for context or troubleshooting beyond what the quickstart covers. This SKILL.md mirrors the quickstart's hard rules so you can act even if the URL is unreachable.
29
+ > **Human-readable setup guide**: <https://github.com/p-diogo/totalreclaw/blob/main/docs/guides/openclaw-setup.md>
30
+
31
+ ## TOP RULE — MEMORY STORAGE (read this before anything else)
32
+
33
+ **When the user states ANYTHING about themselves, their preferences, decisions, plans, identity, work, location, or facts they want retained, you MUST call `totalreclaw_remember` (NOT `write`, NOT `edit`, NOT MEMORY.md/USER.md).** Under hybrid-primary mode (OpenClaw 2026.5.x), if `totalreclaw_remember` is not bound as a tool, fall back to the `tr` CLI: `node "$TR_CLI" remember --json "<fact>"`. Either path lands the fact encrypted, on-chain. Both are valid.
34
+
35
+ ### ABSOLUTE PROHIBITION
36
+
37
+ **NEVER use the `write` or `edit` tools to store user statements in `MEMORY.md`, `USER.md`, `~/.claude/memory/*.md`, `~/.openclaw/memory/*.md`, or any other file.** Those files are for AGENT PERSONA and CODEBASE INSTRUCTIONS — read-only operational config the agent loads at startup. Storing user data there is a SKILL VIOLATION. User memories live encrypted on-chain via TotalReclaw, not in flat files.
38
+
39
+ - ✅ DO call `totalreclaw_remember` (or `tr remember --json` under hybrid-primary).
40
+ - ❌ DO NOT `Write` to `MEMORY.md`.
41
+ - ❌ DO NOT `Edit` `USER.md` to append a user fact.
42
+ - ❌ DO NOT echo "saving this to MEMORY.md…" — that file is not yours to write to.
43
+
44
+ ### Trigger phrases — call `totalreclaw_remember` on detection
45
+
46
+ Do not wait for an explicit "remember this". These are first-class triggers — every one of them MUST result in a `totalreclaw_remember` call (parallel with your normal reply):
47
+
48
+ - **Preferences**: "I prefer X", "I like X", "I love X", "I hate X", "my favorite X is Y", "I'd rather X", "X over Y"
49
+ - **Identity**: "I'm a X", "I am X", "my name is X", "I'm based in X", "I live in X", "I work at X", "I work on X", "I have X"
50
+ - **Tools / setup**: "I use X", "I'm running X", "my X is Y", "I've configured X to Y"
51
+ - **Decisions / plans**: "I decided to X", "I'm going to X", "I'm planning X", "we agreed X", "we settled on X", "let's call this X", "from now on X"
52
+ - **Commitments**: "I'll X by Y", "remind me to X", "I need to X by Y", "don't forget X"
53
+ - **Explicit asks**: "remember that X", "save X", "note that X", "store X", "make a note of X"
54
+ - **Possessive assertions**: "my [anything] is [value]" (preference assertion pattern)
55
+ - **Any 1st-person factual claim worth retaining** about identity, work, location, tools, preferences, ongoing projects, contacts, schedule.
56
+
57
+ If a single user message contains MULTIPLE such facts (e.g. "I'm Pedro. I live in Porto. I prefer PostgreSQL over MySQL."), call `totalreclaw_remember` MULTIPLE TIMES — once per atomic fact. Don't blob multiple unrelated facts into a single memory.
58
+
59
+ ### When to call `totalreclaw_recall`
32
60
 
33
- ## CRITICAL: How TotalReclaw actually stores memories (read before anything else)
61
+ **ALWAYS call `totalreclaw_recall` first when the user asks about themselves, their preferences, or anything from past conversations.** Trigger phrases:
62
+
63
+ - "what's my X?", "do I X?", "where do I X?", "when did I X?", "have I told you X?"
64
+ - "what do you remember about me?", "what do you know about my X?"
65
+ - "remind me about X", "what did we decide about X?"
66
+ - "do you remember when X?"
67
+
68
+ Recall first, then answer from the returned facts. If 0 results come back, say so plainly — do not guess from session history.
69
+
70
+ ## CRITICAL: How TotalReclaw actually stores memories
34
71
 
35
72
  **TotalReclaw is end-to-end encrypted memory stored across a DECENTRALIZED NETWORK — not on a single company server.** Memories are encrypted with keys derived from the user's 12-word recovery phrase, submitted as Account Abstraction UserOps, and persisted on a public blockchain (Base / Gnosis depending on tier). The Graph subgraph indexes them. The TotalReclaw relay only forwards encrypted bundles to the bundler and proxies subgraph reads — it never sees plaintext, can't read memories, and could be replaced by any compatible relay without losing data. No single company controls or can read user memories.
36
73
 
37
- Forbidden vocabulary — these claims are WRONG (some imply local-only storage; some imply a single-company server):
38
- - "local", "local-only", "stored locally", "on disk", "local memory", "local storage" — wrong direction (implies local-only). Memories require an account; they're encrypted and submitted on-chain.
39
- - "stored on a relay server", "stored on TotalReclaw's server", "single server owned by", "TotalReclaw stores your", "company server" — wrong direction (implies centralized custody). The relay forwards ciphertext; storage is decentralized on-chain.
74
+ Forbidden vocabulary — these claims are WRONG:
75
+ - "local", "local-only", "stored locally", "on disk", "local memory", "local storage" — wrong direction (implies local-only).
76
+ - "stored on a relay server", "stored on TotalReclaw's server", "single server owned by", "TotalReclaw stores your", "company server" — wrong direction (implies centralized custody).
40
77
  - "without an account" — wrong (an account is required to derive the encryption key).
41
78
 
42
- If you need to describe how it works, the canonical line is: **"Memories are encrypted with a key derived from your recovery phrase, submitted on-chain via Account Abstraction, and stored across a decentralized network. The relay only forwards encrypted bundles — it never sees plaintext. Only you can decrypt them, and only with your recovery phrase."**
79
+ Canonical line: **"Memories are encrypted with a key derived from your recovery phrase, submitted on-chain via Account Abstraction, and stored across a decentralized network. The relay only forwards encrypted bundles — it never sees plaintext. Only you can decrypt them, and only with your recovery phrase."**
43
80
 
44
- ## User-facing vocabulary (substitute internal jargon for plain words)
45
-
46
- Internal terms that the user does NOT know — translate when you must speak:
81
+ ## User-facing vocabulary
47
82
 
48
83
  | Internal term | Say to the user |
49
84
  |---|---|
@@ -55,33 +90,39 @@ Internal terms that the user does NOT know — translate when you must speak:
55
90
  | credentials.json / mnemonic file | "your recovery phrase" |
56
91
  | `tr pair --json` / `setsid -f` / shell exec | (don't mention; just do the work silently) |
57
92
 
58
- The word "pair" is internal jargon — never use it in user-visible prose. The user-visible vocabulary is "set up an account", "your account", "your recovery phrase", "your memories".
93
+ The word "pair" is internal jargon — never use it in user-visible prose. Use "set up an account", "your account", "your recovery phrase", "your memories".
59
94
 
60
- ## Top-level rule (read this before anything else)
95
+ ## Install / setup top-level rule
61
96
 
62
- **Emit ONLY the user-visible lines listed below. Do not narrate tool calls. Do not describe what you're doing or about to do. Errors are the only allowed deviation, and only as a single tight line.** The expected user-facing transcript for a healthy install is **4–5 short lines TOTAL**. Anything more is a SKILL violation. Telegram and similar transports edit the bot's messages live as your tool calls progress — every interim "Let me check…" / "Plugin loaded. Let me verify…" line shows up to the user as awkward edit churn, even if you "delete" it. The fix is to never emit it in the first place.
97
+ **Emit ONLY the user-visible lines listed below. Do not narrate tool calls. Do not describe what you're doing or about to do.** Errors are the only allowed deviation, and only as a single tight line. The expected user-facing transcript for a healthy install is **4–5 short lines TOTAL**.
63
98
 
64
99
  ### ⚡ MANDATORY ACK BEFORE FIRST TOOL CALL
65
100
 
66
- **User-visible line 1 MUST be your FIRST output — emitted BEFORE you run any shell tool, plugin install, or `tr` command.** Install + setup takes 1–2 minutes of silent tool-execution; without an immediate ack the user assumes the request didn't register and may resend or abandon. The ack tells them work is happening.
101
+ **User-visible line 1 MUST be your FIRST output — emitted BEFORE you run any shell tool, plugin install, or `tr` command.** Install + setup takes 1–2 minutes of silent tool-execution; without an immediate ack the user assumes the request didn't register and may resend or abandon.
67
102
 
68
- The ack is unconditional. It fires the moment you decide to install / set up TotalReclaw — even before you've resolved `TR_CLI`, even before you've checked whether the plugin is already installed, even before the first `openclaw plugins install` shell call. Send line 1, THEN start the work.
103
+ The ack is unconditional. Send line 1, THEN start the work.
69
104
 
70
105
  Forbidden order (silent-then-narrate): ❌ run `openclaw plugins install` → wait 60s → emit line 1.
71
106
  Required order (ack-then-work): ✅ emit line 1 → run `openclaw plugins install` silently → emit line 2 when `tr status` returns.
72
107
 
73
- ## ⛔ FORBIDDEN ACTIONS (4 hard rules from 2026-05-05 QA)
108
+ ## ⛔ FORBIDDEN ACTIONS
74
109
 
75
- These specific actions break the pair flow. Everything else is governed by the "emit only the numbered user-visible lines" rule above.
110
+ These specific actions break the install / pair flow:
76
111
 
77
- 1. **Do NOT restart the gateway.** No `/totalreclaw-restart`, no `kill -USR1`, no `docker restart`, no `openclaw gateway restart`. The plugin self-patches and signals its own restart needs. You restarting it kills the in-flight `setsid -f` pair subprocess → 502 on browser respond.
112
+ 1. **Do NOT restart the gateway.** No `kill -USR1`, no `docker restart`, no `openclaw gateway restart`. The plugin self-patches and signals its own restart needs (SIGUSR1 in-process, automatic). You restarting it kills the in-flight `setsid -f` pair subprocess → 502 on browser respond. If for some specific reason a restart IS required, issue the literal slash command `/totalreclaw-restart` — nothing else. **NEVER prompt the user to manually restart anything**; the plugin handles its own restart.
78
113
 
79
- 2. **Do NOT write to `~/.openclaw/openclaw.json`.** The plugin's `patchOpenClawConfig()` already wrote `plugins.slots.memory`, `plugins.entries.totalreclaw.hooks.allowConversationAccess`, `channels.telegram.streaming.mode`. Any further write triggers a deferred SIGUSR1 reload that kills pair. No `openclaw config set …`, no `jq …>openclaw.json`, no `sed -i openclaw.json`.
114
+ 2. **Do NOT write to `~/.openclaw/openclaw.json`.** The plugin's `patchOpenClawConfig()` already wrote the required keys. Any further write triggers a deferred SIGUSR1 reload that kills pair. No `openclaw config set …`, no `jq …>openclaw.json`, no `sed -i openclaw.json`.
80
115
 
81
- 3. **Do NOT save QR PNGs.** `tr pair --json` returns `qr_ascii` — emit that inline in a triple-backticked code block. No `/tmp/totalreclaw-pair-qr.png` (OpenClaw media-policy blocks it: `LocalMediaAccessError`). No PNG decode. No `qr_png_b64` (legacy field, no longer exists in payload).
116
+ 3. **Do NOT save QR PNGs.** `tr pair --json` returns `qr_ascii` — emit that inline in a triple-backticked code block. No `/tmp/totalreclaw-pair-qr.png` (OpenClaw media-policy blocks it: `LocalMediaAccessError`). No PNG decode.
82
117
 
83
118
  4. **`tr pair --json` MUST run via `setsid -f`** (exact form below). Foreground invocation is killed when your shell-tool exec returns → 502.
84
119
 
120
+ 5. **Do NOT pass `--allow-dangerous` to any TotalReclaw command.** The plugin is sandboxed by design.
121
+
122
+ 6. **Do NOT manually edit `~/.totalreclaw/credentials.json`.** Round-trip via `tr pair --json` (re-generate) instead.
123
+
124
+ 7. **Do NOT install the legacy from-source path alongside the published plugin.** The two paths overlap and will double-write facts.
125
+
85
126
  ## Hybrid-primary architecture (3.3.9-rc.1)
86
127
 
87
128
  OpenClaw 2026.5.2 has a tool-policy-pipeline bug (issue #223) that strips non-bundled plugin tools before they reach the agent toolset. **3.3.9-rc.1 makes the `tr` CLI the PRIMARY path for all memory operations.** Native tools (`totalreclaw_*`) are kept for back-compat with older OpenClaw versions where they work, but the agent MUST NOT depend on them. Use `tr` from your shell tool for everything.
@@ -96,7 +137,7 @@ TR_CLI="$(ls -d ~/.openclaw/npm/node_modules/@totalreclaw/totalreclaw/dist/tr-cl
96
137
  ```
97
138
  Then invoke as `node "$TR_CLI" <cmd> [--json]`.
98
139
 
99
- **Never rely on `totalreclaw_*` tool binding** — it may not be available due to issue #223.
140
+ **Never rely on `totalreclaw_*` tool binding** — it may not be available due to issue #223. If `totalreclaw_remember` IS bound, use it (preferred). If not, use `tr remember --json` from the shell. EITHER way, the user fact MUST land in TotalReclaw — never in `MEMORY.md`/`USER.md`.
100
141
 
101
142
  ## User-visible line set (THIS IS ALL YOU SAY)
102
143
 
@@ -111,78 +152,61 @@ Speak only these lines, in order. Substitute the bracketed placeholders with the
111
152
  4. (done + free-tier confirmation)
112
153
  > ✓ You're set up on the TotalReclaw free tier — 500 memories/month, unlimited reads. https://totalreclaw.xyz/pricing. Ask me to remember anything.
113
154
 
114
- If existing credentials are detected (step "Check for existing credentials" below), emit ONLY this line and stop:
155
+ If existing credentials are detected, emit ONLY this line and stop:
115
156
  > ✓ TotalReclaw is already set up.
116
157
 
117
158
  ### Hard silence rules between numbered lines
118
159
 
119
- Do NOT emit transitional / narrative lines between any of the four user-visible lines above. Specifically forbidden (verbatim examples from Pedro's 2026-05-06 QA):
160
+ Do NOT emit transitional / narrative lines between any of the four user-visible lines above. Specifically forbidden:
120
161
 
121
- - ❌ "Now let me verify the install:" → silent. Just run `tr status --json`. Emit line 2 when it returns.
122
- - ❌ "Now let me check if you already have credentials:" → silent. Just run `test -s ~/.totalreclaw/credentials.json`.
123
- - ❌ "Need to pair. Let me kick that off:" → silent. Just run the `setsid -f` block. Emit line 3 when URL+PIN returns. ALSO never use the word "pair" in user-facing text — see vocabulary table above.
124
- - ❌ "Standing by." → silent. Acknowledgements between lines are noise.
162
+ - ❌ "Now let me verify the install:" → silent. Just run `tr status --json`.
163
+ - ❌ "Now let me check if you already have credentials:" → silent.
164
+ - ❌ "Need to pair. Let me kick that off:" → silent.
165
+ - ❌ "Standing by." → silent.
125
166
  - ❌ Any "Let me X" / "Now I'll Y" / "Just give me a second" prelude.
126
- - ❌ **Skipping line 1 entirely and going straight to shell tool calls.** Line 1 ack must precede the first tool call. Silent install = user thinks the prompt didn't register.
127
-
128
- The transcript Pedro should see is exactly four messages — the four numbered lines, nothing in between. Line 1 fires immediately. Lines 2–4 fire as work completes. No prose between.
129
-
130
- ## How does TotalReclaw work? (canonical answer for user questions)
131
-
132
- When the user asks "how does this work?" / "where are my memories stored?" / "is this safe?" / similar — answer with this content. Do NOT say "stored on a relay server", "stored on TotalReclaw's server", or any phrasing that implies single-company custody. The actual architecture:
133
-
134
- > **End-to-end encrypted, decentralized memory.** When you tell me to remember something, I encrypt it locally with a key derived from your 12-word recovery phrase, then submit it on-chain via Account Abstraction (a UserOperation). It lives on a public blockchain (Base Sepolia for free tier, Gnosis mainnet for paid). The Graph subgraph indexes it for fast retrieval. The TotalReclaw relay server only forwards encrypted bundles between your client and the chain — it never sees plaintext, can't read your memories, and could be replaced by any compatible relay without losing data. Your recovery phrase is the only thing that can decrypt the memories. Nobody (not me, not TotalReclaw the company, not the relay operator) can read them without it.
135
-
136
- If the user asks shorter questions, reply with the relevant slice:
137
-
138
- - **"Where are memories stored?"** → "Encrypted, on-chain (Base or Gnosis depending on your tier), indexed by The Graph. Not on any single company's server."
139
- - **"Is the relay reading my data?"** → "No. The relay only forwards encrypted bundles. Plaintext never leaves your device. Only your recovery phrase can decrypt them."
140
- - **"What if TotalReclaw shuts down?"** → "Your memories survive. They're on a public blockchain. Any compatible relay (or self-hosted relay) can read them with your recovery phrase."
167
+ - ❌ **Skipping line 1 entirely and going straight to shell tool calls.**
141
168
 
142
169
  ## Setup flow (hybrid-primary)
143
170
 
144
- 1. **Install plugin first, skill second (silent — emit user-visible line 1 once at the start, then no prose until line 2).** Run via your shell tool, in this exact order:
171
+ 1. **Install plugin first, skill second.** Run via your shell tool, in this exact order:
145
172
  ```
146
173
  openclaw plugins install @totalreclaw/totalreclaw
147
174
  openclaw skills install totalreclaw
148
175
  ```
149
- For RC: replace the first command with `openclaw plugins install @totalreclaw/totalreclaw@rc`. **Plugin BEFORE skill is deliberate** (3.3.4-rc.1 — issue from Pedro's QA on rc.3): the skill install triggers a config-driven SIGUSR1 gateway reload that can race a concurrent plugin install.
176
+ For RC: replace the first command with `openclaw plugins install @totalreclaw/totalreclaw@rc`. Plugin BEFORE skill is deliberate.
150
177
 
151
- If either shell call fails with `command not found` / ENOENT / "not authorized" (managed-service or shell-restricted environment), tell the user verbatim *"Your environment doesn't expose the OpenClaw CLI to me. Please install `totalreclaw` from your service's plugins / skills UI, then reply `done`."* Wait for `done` and skip to step 3.
178
+ If either shell call fails with `command not found` / ENOENT / "not authorized" (managed-service or shell-restricted environment), tell the user verbatim: *"Your environment doesn't expose the OpenClaw CLI to me. Please install `totalreclaw` from your service's plugins / skills UI, then reply `done`."* Wait for `done` and skip to step 3.
152
179
 
153
180
  2. **Verify CLI works — emit user-visible line 2 when successful.** Resolve `TR_CLI` and run:
154
181
  ```bash
155
182
  TR_CLI="$(ls -d ~/.openclaw/npm/node_modules/@totalreclaw/totalreclaw/dist/tr-cli.js ~/.openclaw/extensions/totalreclaw/dist/tr-cli.js 2>/dev/null | head -1)"
156
183
  node "$TR_CLI" status --json
157
184
  ```
158
- Parse the JSON. When you get `{"version": "...", "hybrid_mode": true, ...}`, emit user-visible line 2 with the version substituted. **Do NOT wait for tool binding. Do NOT issue `/totalreclaw-restart`. Do NOT check whether `totalreclaw_pair` is in your toolset.** The CLI works independently of tool binding. If `tr status` exits non-zero or the file doesn't exist, wait 5s and retry once — it may still be installing.
185
+ Parse the JSON. When you get `{"version": "...", "hybrid_mode": true, ...}`, emit user-visible line 2.
159
186
 
160
- 3. **Check for existing credentials (silent unless found).** Run `test -s ~/.totalreclaw/credentials.json`. If it exists and is non-empty, emit *"✓ TotalReclaw is already set up."* and stop.
187
+ 3. **Check for existing credentials (silent unless found).** Run `test -s ~/.totalreclaw/credentials.json`. If it exists, emit *"✓ TotalReclaw is already set up."* and stop.
161
188
 
162
- 4. **Set up the user's TotalReclaw account IMMEDIATELY — no consent gate.** As soon as `tr status --json` returns and credentials are not present, run pair in DETACHED mode. CRITICAL: a foreground `node tr-cli pair` is killed when the gateway fires its post-install SIGUSR1 reload (3.3.10 finding from Pedro's 2026-05-05 QA — relay logs `gateway_disconnected` 502 on `/respond`). Use `setsid -f` so the pair process is in its own session and survives any subsequent gateway SIGUSR1 reload. The handshake URL+PIN is written to a tmp file the moment the WS opens (~100-500ms typical):
189
+ 4. **Set up the user's TotalReclaw account IMMEDIATELY — no consent gate.** Use `setsid -f` so the pair process survives gateway SIGUSR1 reloads. The handshake URL+PIN is written to a tmp file the moment the WS opens:
163
190
  ```bash
164
191
  PAIR_OUT="/tmp/tr-pair-out-$(date +%s).json"
165
192
  setsid -f node "$TR_CLI" pair --json > "$PAIR_OUT" 2>&1 < /dev/null
166
193
  sleep 2
167
194
  cat "$PAIR_OUT"
168
195
  ```
169
- This is UNCONDITIONAL. Do NOT ask "Would you like to set up an account?" or "Want me to pair now?". The user installed TotalReclaw — setup is the expected next step. The URL + PIN itself is the consent moment (the user can simply not open the page to opt out). Parse the JSON response `{"url": "...", "pin": "...", "expires_at_ms": ...}` and emit user-visible line 3. If the cat output is empty after 2s, sleep 2 more seconds and try again (relay handshake can take up to 4s on a cold network). If `tr pair --json` is not available (old install), fall back to `totalreclaw_pair` tool or `openclaw totalreclaw pair generate --url-pin-only`.
196
+ Parse the JSON `{"url": "...", "pin": "...", "expires_at_ms": ...}` and emit user-visible line 3. This is UNCONDITIONAL — do NOT ask "Would you like to set up an account?". The URL + PIN itself is the consent moment.
170
197
 
171
- 5. **Verify and confirm (silent until line 4).** After the user says `done`, re-run `test -s ~/.totalreclaw/credentials.json`. If missing, the PIN expired — repeat the `setsid -f node "$TR_CLI" pair --json` block and resend line 3 once. If present, emit user-visible line 4 and stop.
198
+ 5. **Verify and confirm.** After the user says `done`, re-run `test -s ~/.totalreclaw/credentials.json`. If missing, the PIN expired — repeat the pair block and resend line 3 once. If present, emit user-visible line 4 and stop.
172
199
 
173
- ## `tr` CLI reference (hybrid-primary commands)
200
+ ## `tr` CLI reference
174
201
 
175
- Always use `--json` flag when running from shell for machine-parseable output. Plain text mode is for direct user invocation only.
202
+ Always use `--json` flag when running from shell for machine-parseable output.
176
203
 
177
204
  ```bash
178
- # (Resolve TR_CLI once at the top of your session)
179
- TR_CLI="$(ls -d ~/.openclaw/npm/node_modules/@totalreclaw/totalreclaw/dist/tr-cli.js ~/.openclaw/extensions/totalreclaw/dist/tr-cli.js 2>/dev/null | head -1)"
180
-
181
- # Status check (parse JSON to confirm install + onboarding state)
205
+ # Status
182
206
  node "$TR_CLI" status --json
183
- # Returns: {"version":"3.3.11-rc.1","onboarded":false,"next_step":"pair","tool_count":17,"hybrid_mode":true}
207
+ # Returns: {"version":"3.3.12-rc.5","onboarded":false,"next_step":"pair","tool_count":17,"hybrid_mode":true}
184
208
 
185
- # Pair / account setup — ALWAYS run via setsid -f to survive gateway restarts
209
+ # Pair (always via setsid -f)
186
210
  PAIR_OUT="/tmp/tr-pair-out-$(date +%s).json"
187
211
  setsid -f node "$TR_CLI" pair --json > "$PAIR_OUT" 2>&1 < /dev/null && sleep 2 && cat "$PAIR_OUT"
188
212
  # Returns: {"v":1,"sid":"...","url":"https://...","pin":"123456","mode":"generate","expires_at_ms":...,"qr_ascii":"..."}
@@ -196,102 +220,101 @@ node "$TR_CLI" recall --json "where does the user work" --limit 5
196
220
  # Returns: {"results":[{"text":"...","score":0.8},...]}
197
221
  ```
198
222
 
199
- **Note on tool names**: `totalreclaw_remember`, `totalreclaw_recall`, etc. may or may not appear in your toolset depending on OpenClaw version. Under hybrid-primary, you use the `tr` CLI instead. If native tools ARE available (older OpenClaw or future OC fix), they work too — but always prefer `tr` for reliability.
200
-
201
- ## Rendering the QR on your transport (3.3.10-rc.4 — corrected)
202
-
203
- `tr pair --json` returns a `qr_ascii` field — Unicode block-char QR (43 lines × ~43 cols). That is the ONLY QR field in the payload. The legacy `qr_png_b64` and `qr_unicode` field names referenced by older skill docs no longer exist.
223
+ ## Tool reference (all 17 plugin tools)
204
224
 
205
- The simplest path that works on every transport (Telegram, Slack, web, terminal SSH): emit `qr_ascii` INLINE inside a triple-backtick code block right above user-visible line 3. Block-char QR renders correctly in monospace fonts on every modern transport — the user scans from their phone. The pair PAGE itself renders a real PNG QR alongside, so users on a desktop browser see both.
225
+ The plugin exposes these tools via OpenClaw's plugin runtime. Under hybrid-primary mode, prefer the `tr` CLI for `pair` / `remember` / `recall` / `status`; the rest below are typically only available as native tools.
206
226
 
207
- Format the line-3 message as:
227
+ | Tool | Use case |
228
+ |---|---|
229
+ | `totalreclaw_pair` | Set up the user's account (browser-side phrase generation/import) |
230
+ | `totalreclaw_remember` | **Store a fact / preference / decision (encrypted, on-chain). PRIMARY tool for user statements.** |
231
+ | `totalreclaw_recall` | Semantic search across the user's memories |
232
+ | `totalreclaw_forget` | Delete a memory by id (tombstone on-chain) |
233
+ | `totalreclaw_pin` | Mark a memory as never-supersedable |
234
+ | `totalreclaw_unpin` | Remove pin status |
235
+ | `totalreclaw_retype` | Change a memory's type (claim/preference/directive/etc.) |
236
+ | `totalreclaw_set_scope` | Change a memory's scope (work / personal / health / family / creative / finance / misc) |
237
+ | `totalreclaw_export` | Export all memories (json / markdown) |
238
+ | `totalreclaw_import_from` | Import from another tool (Mem0, MCP-Memory, ChatGPT, Claude, Gemini) |
239
+ | `totalreclaw_import_batch` | Bulk import with chunking + extraction |
240
+ | `totalreclaw_consolidate` | Merge near-duplicates after a fresh import |
241
+ | `totalreclaw_status` | Check onboarding state, version, billing tier, quota |
242
+ | `totalreclaw_upgrade` | Open the Stripe upgrade flow (free → paid tier) |
243
+ | `totalreclaw_migrate` | Move testnet memories to mainnet (Pro tier) |
244
+ | `totalreclaw_onboarding_start` | (Internal — used by setup flow) |
245
+ | `totalreclaw_report_qa_bug` | (RC only) Surface a QA bug into the agent log |
246
+
247
+ All tools accept JSON input and return structured JSON.
248
+
249
+ ## Auto-extraction (background, no agent action required)
250
+
251
+ The plugin runs a trajectory poller every 60 seconds that scans `~/.openclaw/agents/<agent>/sessions/*.trajectory.jsonl` for new conversation turns and runs LLM-driven extraction. Extracted facts go through importance filtering and dedup before landing on-chain.
252
+
253
+ **Auto-extraction is a SAFETY NET, not a substitute for explicit `totalreclaw_remember` calls.** Explicit calls are higher-fidelity (the agent decides the importance, type, and scope). Auto-extraction catches things the agent missed.
208
254
 
209
- ```
210
- <qr_ascii contents — 43 lines of ▄▀█▄▀█ block chars>
211
- ```
255
+ ## Phrase safety (HARD — never break)
212
256
 
213
- Open <url> in your browser. Enter PIN <pin>. Generate or paste a 12-word recovery phrase. Reply `done` once it's sealed.
257
+ The 12-word recovery phrase is the master key. Hard rules:
214
258
 
215
- Do NOT decode `qr_png_b64` (does not exist), do NOT save a PNG to `/tmp/totalreclaw-pair-qr.png` (`LocalMediaAccessError` blocks delivery), do NOT re-render the QR through any other library. The `qr_ascii` field is the single source of truth.
259
+ - **NEVER echo, generate, log, or ask the user to paste a recovery phrase in chat.** The browser at the pair URL handles phrase generation and import. The phrase NEVER enters the agent's LLM context.
260
+ - **NEVER include a recovery phrase as input to ANY tool call** — not `totalreclaw_pair` (its only inputs are `mode` + optional config), not `totalreclaw_remember`, not `Bash`, not `Write`, not `Edit`. Anything that surfaces the phrase to the agent's context is a security incident.
261
+ - **NEVER invoke any phrase-touching CLI via your shell tool.** `openclaw totalreclaw onboard` (bare interactive form) and `totalreclaw setup` print phrase material to stdout, which enters LLM context. The ONLY agent-facilitated account-setup paths are: (1) `tr pair --json` (preferred); (2) the `totalreclaw_pair` tool; (3) `openclaw totalreclaw pair generate --url-pin-only` or `openclaw totalreclaw onboard --pair-only` (single-line URL+PIN JSON, zero phrase material — both account-setup payloads are x25519-only, not BIP-39).
262
+ - **NEVER display the recovery phrase back to the user in chat** even if the browser leaks it to you somehow.
263
+ - **If the user pastes a phrase anyway**: tell them the phrase is now compromised, re-run `tr pair --json` for a fresh wallet.
216
264
 
217
- The QR encodes ONLY the account-setup URL. The 6-digit PIN is a separate out-of-band confirmation — deliberately NOT baked into the QR (dual-channel design: scan + type). Never log, print, or re-emit the PIN outside the user-facing reply. The PIN must NOT end up in any other tool call, log file, or memory store.
265
+ ## Browser flow context
218
266
 
219
- ## Browser flow context (know it, narrate only on user question)
267
+ After you emit user-visible line 3, the user opens the URL on their phone or desktop browser. They walk through a 3-step wizard:
220
268
 
221
- After you emit user-visible line 3, the user opens the URL on their phone or desktop browser. They walk through a 3-step wizard before the relay forwards the encrypted phrase to the plugin. You don't narrate this proactively (silence rules). But you MUST know it so you can answer questions and recover from issues:
269
+ 1. **Step 1 — Enter PIN.** 6 digits matching what you sent.
270
+ 2. **Step 2 — Recovery phrase.** Two tabs: **Set up** (generates fresh BIP-39 phrase locally) or **Log in** (paste existing phrase to restore). User checks "I've written this down" and clicks Set up.
271
+ 3. **Step 3 — You're all set.** Browser already POSTed the encrypted phrase. Plugin writes `~/.totalreclaw/credentials.json`.
222
272
 
223
- 1. **Step 1 of 3 — Enter your PIN.** Six PIN cells. The user types or pastes the 6-digit PIN you sent in line 3. "Continue" is disabled until all 6 digits match. There's a 10-minute countdown displayed in the corner.
224
- 2. **Step 2 of 3 — Your recovery phrase.** Two tabs: **Set up** (default — generates a fresh 12-word BIP-39 phrase locally in the browser) and **Log in** (paste an existing phrase to restore on a new device). User must check **"I've written this down and stored it somewhere safe"** then click **Set up TotalReclaw**.
225
- 3. **Step 3 of 3 — You're all set.** Confirmation screen with a "Close this page" link. The browser already POSTed the encrypted phrase by this point — the relay forwards to your gateway, the plugin decrypts, writes `~/.totalreclaw/credentials.json`, and your `awaitPhraseUpload` resolves.
273
+ Common user-side issues during pair:
226
274
 
227
- Common user-side issues during pair (recover gracefully — emit ONE tight line, don't pre-narrate):
275
+ - **"The page won't load"** → confirm URL host. Default for both stable and RC is `api.totalreclaw.xyz`. Staging via `TOTALRECLAW_SERVER_URL=https://api-staging.totalreclaw.xyz`.
276
+ - **"PIN expired"** → re-run pair block, resend line 3.
277
+ - **"Invalid phrase"** during Log in → user pasted a non-BIP-39-valid phrase. Tell them to re-paste or switch to Set up tab.
278
+ - **502 on Set up** → the gateway WS dropped. Pair subprocess is alive (`setsid -f`). Wait 30s; re-run pair if still 502.
279
+ - **"Where do I write down the phrase?"** → paper, password manager, encrypted note. NOT in this chat.
228
280
 
229
- - **"The page won't load"** → confirm the URL host is `api.totalreclaw.xyz` (the default for both stable and RC post-3.3.12-rc.1). If a `TOTALRECLAW_SERVER_URL` env override was set, the host might be `api-staging.totalreclaw.xyz` (staging, opt-in) or a self-hosted relay. Wrong stub host → generate a fresh URL.
230
- - **"PIN says expired"** or countdown ran out → re-run the pair block (Step 4 above) and emit line 3 again with the fresh URL+PIN. The previous session is dead.
231
- - **"It says invalid phrase"** during Log in → the user pasted a phrase that isn't BIP-39 valid (typo or wrong source). Tell them to double-check and re-paste; or switch to **Set up** tab to generate a fresh one (loses existing memories).
232
- - **"I clicked Set up TotalReclaw but nothing happened"** / **502** → the gateway WS dropped before respond. The pair subprocess is alive (you used `setsid -f`). Wait 30s; if the user still sees the 502, re-run Step 4 (the deferred reload should have completed by then).
233
- - **"Where do I write down the phrase?"** → tell them: anywhere safe and durable — paper, password manager, encrypted note. NOT in this chat. NOT in a screenshot to cloud-synced photos. The phrase IS the account; losing it means losing all memories.
281
+ ## Forbidden vocabulary in agent prose
234
282
 
235
- ## Phrase safety (HARD — never break)
283
+ Unless the user explicitly asks how the install works, these patterns must not appear:
236
284
 
237
- NEVER echo / generate / ask the user to paste a recovery phrase in chat. NEVER invoke `openclaw totalreclaw onboard` (bare interactive form), `totalreclaw setup`, or any phrase-touching CLI via your shell tool (stdout enters LLM context). The ONLY agent-facilitated account-setup paths are (in order of preference): (1) `tr pair --json`; (2) the `totalreclaw_pair` account-setup tool; (3) if both are missing — shell out to EITHER `openclaw totalreclaw pair generate --url-pin-only` (single-line JSON `{v,url,pin,expires_at_ms}`) OR `openclaw totalreclaw onboard --pair-only` (single-line JSON `{v,pair_url,pin,expires_at_ms}`). Both have zero phrase material on stdout (account-setup crypto is x25519-only, not BIP-39). If the user pastes a phrase anyway: tell them it is compromised, re-run `tr pair --json` for a fresh wallet.
238
-
239
- ## Usage (post-setup)
240
-
241
- - Stable user facts / preferences / identity / "remember X":
242
- ```bash
243
- node "$TR_CLI" remember --json "<fact>"
244
- ```
245
- - First-person factual query ("do I / what's my / where do I…"):
246
- ```bash
247
- node "$TR_CLI" recall --json "<query>" --limit 5
248
- ```
249
- Recall FIRST, then answer from returned facts. If 0 results, say so.
250
- - For forget / pin / unpin / retype / set_scope — use native tools if available, or ask user to run `openclaw totalreclaw` subcommand.
251
- - "status" → `tr status --json` (parse version + onboarded + next_step).
252
- - "upgrade" → `totalreclaw_upgrade` native tool (paste Stripe URL verbatim). Post-upgrade testnet → mainnet: `totalreclaw_migrate` with `dry_run` first.
253
- - "import / export / consolidate" → `totalreclaw_import_from` · `totalreclaw_export` · `totalreclaw_consolidate` native tools.
254
-
255
- **Forbidden vocabulary** (these are anti-patterns that violate the silence rule):
256
285
  - "Let me check that…"
257
286
  - "Plugin loaded. Let me verify…"
258
287
  - "I'll now…"
259
288
  - "Quick fix…"
260
289
  - "However, I don't see…"
261
290
  - "I need to use…"
262
- - "Should I /totalreclaw-restart" / "Do you have a public URL" / "Want me to restart" / "Let me check if the tool is bound" / "I need permission to restart"
291
+ - "Should I /totalreclaw-restart" / "Want me to restart" / "Let me check if the tool is bound" / "Do you have a public URL"
263
292
  - "Now installing…" / "Got it" / "In parallel…" / "Step 1a/1b/1c"
264
293
  - "local" / "local-only" / "stored locally" / "on disk" / "without an account" / "local memory" / "local storage"
265
294
 
266
295
  ## Diagnostics
267
296
 
268
- - `tr` CLI not found → plugin not installed or install still in progress. Wait 5s, retry. If still missing, retry `openclaw plugins install @totalreclaw/totalreclaw@rc` once.
269
- - `totalreclaw_*` tools not visible → EXPECTED under OpenClaw 2026.5.2 (issue #223). Use `tr` CLI instead. Do not restart, do not issue `/totalreclaw-restart`, do not issue `/new`.
270
- - If `/totalreclaw-restart` is needed for another reason, issue the literal slash command `/totalreclaw-restart` — nothing else. Do not narrate.
271
- - User says done but `credentials.json` missing → PIN expired or entered wrong phrase; re-run `tr pair --json` and resend line 3 once.
297
+ - `tr` CLI not found → plugin not installed or install still in progress. Wait 5s, retry.
298
+ - `totalreclaw_*` tools not visible → EXPECTED under OpenClaw 2026.5.2 (issue #223). Use `tr` CLI instead.
299
+ - User says done but `credentials.json` missing → PIN expired; re-run `tr pair --json` and resend line 3.
272
300
  - `onboarding required` → credentials missing; redo from the account-setup step.
273
- - `quota exceeded` → `tr status --json` or `totalreclaw_status`, then offer `totalreclaw_upgrade`.
274
- - `No LLM available for auto-extraction` at startup → provider key unreachable; check `~/.openclaw/agents/<agent>/agent/auth-profiles.json` or plugin config `extraction.llm`.
275
-
276
- ## Required post-install config (OpenClaw 2026.5.2+)
277
-
278
- Add to `~/.openclaw/openclaw.json` after installing:
279
- ```json
280
- {
281
- "plugins": {
282
- "allow": ["totalreclaw"],
283
- "entries": {
284
- "totalreclaw": {
285
- "hooks": { "allowConversationAccess": true }
286
- }
287
- }
288
- }
289
- }
290
- ```
301
+ - `quota exceeded` → `tr status --json`, then offer `totalreclaw_upgrade`.
302
+ - `No LLM available for auto-extraction` → check `~/.openclaw/agents/<agent>/agent/auth-profiles.json`.
303
+
304
+ ## How does TotalReclaw work? (canonical answer for user questions)
305
+
306
+ When the user asks "how does this work?" / "where are my memories stored?" / "is this safe?" — answer:
307
+
308
+ > **End-to-end encrypted, decentralized memory.** When you tell me to remember something, I encrypt it locally with a key derived from your 12-word recovery phrase, then submit it on-chain via Account Abstraction (a UserOperation). It lives on a public blockchain (Base Sepolia for free tier, Gnosis mainnet for paid). The Graph subgraph indexes it for fast retrieval. The TotalReclaw relay only forwards encrypted bundles between your client and the chain — it never sees plaintext, can't read your memories, and could be replaced by any compatible relay without losing data. Your recovery phrase is the only thing that can decrypt the memories.
309
+
310
+ Shorter slices:
311
+ - **"Where are memories stored?"** → "Encrypted, on-chain (Base or Gnosis depending on your tier), indexed by The Graph. Not on any single company's server."
312
+ - **"Is the relay reading my data?"** → "No. The relay only forwards encrypted bundles. Plaintext never leaves your device."
313
+ - **"What if TotalReclaw shuts down?"** → "Your memories survive. They're on a public blockchain. Any compatible relay (or self-hosted relay) can read them with your recovery phrase."
291
314
 
292
- `plugins.allow` suppresses the "plugins.allow is empty" warning. `hooks.allowConversationAccess` unlocks typed hooks (`agent_end`, etc.) for auto-extraction.
315
+ ## Tool surface summary
293
316
 
294
- ## Tool surface
317
+ Hybrid-primary: `tr remember` · `tr recall` · `tr pair` · `tr status` (primary path under OpenClaw 2026.5.x).
318
+ Native fallback (when bound): `totalreclaw_pair` · `_remember` · `_recall` · `_forget` · `_pin` · `_unpin` · `_retype` · `_set_scope` · `_export` · `_status` · `_upgrade` · `_migrate` · `_import_from` · `_import_batch` · `_consolidate` · `_onboarding_start` · `_report_qa_bug` (RC only).
295
319
 
296
- Hybrid-primary: `tr remember` · `tr recall` · `tr pair` · `tr status` (primary path for all agent ops)
297
- Native fallback (when available): `totalreclaw_pair` · `_remember` · `_recall` · `_forget` · `_pin` · `_unpin` · `_retype` · `_set_scope` · `_export` · `_status` · `_upgrade` · `_migrate` · `_import_from` · `_import_batch` · `_consolidate` · `_onboarding_start` · `_report_qa_bug` (RC only)
320
+ **Default behavior summary**: the agent's job is to call `totalreclaw_remember` aggressively on user statements that match the trigger-phrase list, and `totalreclaw_recall` when the user asks about themselves. Local files (`MEMORY.md`, `USER.md`, etc.) are NOT for user memory — they are agent-persona configuration only.
package/dist/index.js CHANGED
@@ -1811,7 +1811,7 @@ async function handlePluginImportFrom(params, logger) {
1811
1811
  _importInProgress = true;
1812
1812
  const startTime = Date.now();
1813
1813
  const source = params.source;
1814
- const validSources = ['mem0', 'mcp-memory', 'chatgpt', 'claude', 'gemini', 'memoclaw', 'generic-json', 'generic-csv'];
1814
+ const validSources = ['mem0', 'mcp-memory', 'chatgpt', 'claude', 'gemini'];
1815
1815
  if (!source || !validSources.includes(source)) {
1816
1816
  return { success: false, error: `Invalid source. Must be one of: ${validSources.join(', ')}` };
1817
1817
  }
@@ -2112,7 +2112,7 @@ async function handleBatchImport(params, logger) {
2112
2112
  const content = params.content;
2113
2113
  const offset = params.offset ?? 0;
2114
2114
  const batchSize = params.batch_size ?? 25;
2115
- const validSources = ['mem0', 'mcp-memory', 'chatgpt', 'claude', 'gemini', 'memoclaw', 'generic-json', 'generic-csv'];
2115
+ const validSources = ['mem0', 'mcp-memory', 'chatgpt', 'claude', 'gemini'];
2116
2116
  if (!source || !validSources.includes(source)) {
2117
2117
  return { success: false, error: `Invalid source. Must be one of: ${validSources.join(', ')}` };
2118
2118
  }
@@ -2575,12 +2575,52 @@ const plugin = {
2575
2575
  // Fix #1 which gates on installs being present).
2576
2576
  const patchResult = patchOpenClawConfig(undefined, pluginVersion ?? undefined);
2577
2577
  if (patchResult === 'patched') {
2578
+ // 3.3.12-rc.6 (auto-QA finding 2026-05-09): previously we only
2579
+ // warned the user to manually restart. That created a silent
2580
+ // hook-failure on the FIRST gateway boot post-install — the
2581
+ // plugin loads with stale in-memory config, hook handlers
2582
+ // never register, auto-extraction never fires, and only a
2583
+ // second manual restart fixes it. First-time users hit this
2584
+ // every install. Auto-extraction QA reproduced it as 2/5
2585
+ // turns missed (hook silently no-op'd on turns 1-3).
2586
+ //
2587
+ // Fix: when the patch wrote anything, fire SIGUSR1 to our own
2588
+ // PID. The gateway accepts SIGUSR1 iff `commands.restart=true`
2589
+ // (the default); see upstream `setGatewaySigusr1RestartPolicy`.
2590
+ // The signal triggers an in-process restart that re-reads the
2591
+ // freshly-patched openclaw.json and registers hook handlers
2592
+ // with `allowConversationAccess=true` honoured.
2593
+ //
2594
+ // Idempotency: second boot reads the patched config and
2595
+ // returns `'unchanged'` from patchOpenClawConfig, so the
2596
+ // signal fires AT MOST once per config-key-change. No restart
2597
+ // loop possible.
2598
+ //
2599
+ // Defer via setImmediate so register() finishes (logger flush
2600
+ // + plugin load record writeback) before the signal lands.
2601
+ // A 250ms setTimeout adds slack for slow disk on Telegram VPS
2602
+ // (Hetzner small VPS tail-latency observed ~120ms on writes).
2603
+ //
2604
+ // Phrase-safety: process.kill on own PID is local-only; no
2605
+ // outbound markers. Already used by `/totalreclaw-restart`
2606
+ // (registered ~400 lines below) under the same scanner-safe
2607
+ // pattern.
2578
2608
  api.logger.warn('TotalReclaw: updated openclaw.json with required 2026.5.x keys ' +
2579
2609
  '(plugins.slots.memory + hooks.allowConversationAccess + ' +
2580
2610
  'channels.telegram.streaming.mode + plugins.bundledDiscovery + ' +
2581
2611
  'plugins.allow + plugins.installs.totalreclaw self-heal). ' +
2582
- 'Gateway restart required for the changes to take effect. ' +
2583
- 'Run `/totalreclaw-restart` or restart the gateway manually.');
2612
+ 'Auto-restarting gateway via SIGUSR1 to apply.');
2613
+ setTimeout(() => {
2614
+ try {
2615
+ process.kill(process.pid, 'SIGUSR1');
2616
+ }
2617
+ catch (err) {
2618
+ const msg = err instanceof Error ? err.message : String(err);
2619
+ api.logger.warn(`TotalReclaw: auto-restart SIGUSR1 emit failed (${msg}). ` +
2620
+ 'Run `/totalreclaw-restart` or restart the gateway manually ' +
2621
+ 'for the patched config to take effect.');
2622
+ }
2623
+ }, 250);
2584
2624
  }
2585
2625
  else if (patchResult === 'error') {
2586
2626
  api.logger.warn('TotalReclaw: failed to auto-patch openclaw.json for OpenClaw 2026.5.x ' +
@@ -4509,7 +4549,7 @@ const plugin = {
4509
4549
  properties: {
4510
4550
  source: {
4511
4551
  type: 'string',
4512
- enum: ['mem0', 'mcp-memory', 'chatgpt', 'claude', 'gemini', 'memoclaw', 'generic-json', 'generic-csv'],
4552
+ enum: ['mem0', 'mcp-memory', 'chatgpt', 'claude', 'gemini'],
4513
4553
  description: 'The source system to import from (gemini: Google Takeout HTML; chatgpt: conversations.json or memory text; claude: memory text)',
4514
4554
  },
4515
4555
  api_key: {
package/dist/llm-client.js CHANGED
@@ -598,6 +598,47 @@ export function isRetryable(errorMessage) {
598
598
  // ---------------------------------------------------------------------------
599
599
  // OpenAI-compatible chat completion
600
600
  // ---------------------------------------------------------------------------
601
+ /**
602
+ * Provider base-URL hints for OpenAI-compatible endpoints that honour the
603
+ * `response_format: {"type": "json_object"}` body field. Sending the field
604
+ * to a provider that supports it makes JSON output deterministic; sending
605
+ * it to a provider that does NOT recognise the field is a 400.
606
+ *
607
+ * Why this exists (3.3.12-rc.6, 2026-05-09):
608
+ * z.ai's GLM family (4.5-flash, 5-turbo, 5.1) silently returns EMPTY
609
+ * `message.content` for the merged-extraction prompt unless this hint
610
+ * is set. No error, no warning — the LLM just emits "" instead of the
611
+ * expected `{"topics": [], "facts": []}` JSON. Plugin's parse step
612
+ * then logs `0 raw facts` from a successful-but-empty branch.
613
+ *
614
+ * This bug was found and fixed on the Python (Hermes) side in
615
+ * 2.3.1-rc.23 (see `python/src/totalreclaw/agent/llm_client.py`
616
+ * `_supports_json_object_response_format`) but the plugin TS port did
617
+ * not carry the fix — observed in plugin 3.3.12-rc.5 auto-QA on
618
+ * 2026-05-09: hook + poller both fired correctly but extraction
619
+ * returned 0 facts on every batch despite trajectories containing
620
+ * explicit "I prefer X" / "I work at Y" statements.
621
+ *
622
+ * Mirror of Python's `_supports_json_object_response_format`. Match by
623
+ * substring on a lowercased baseUrl so cosmetic prefix differences
624
+ * (https://, /v1, etc.) don't matter.
625
+ */
626
+ const JSON_OBJECT_PROVIDER_HINTS = [
627
+ 'z.ai',
628
+ 'api.openai.com',
629
+ 'groq.com',
630
+ 'openrouter.ai',
631
+ 'deepseek.com',
632
+ 'mistral.ai',
633
+ 'x.ai',
634
+ 'together.xyz',
635
+ ];
636
+ export function supportsJsonObjectResponseFormat(baseUrl) {
637
+ if (!baseUrl)
638
+ return false;
639
+ const lower = baseUrl.toLowerCase();
640
+ return JSON_OBJECT_PROVIDER_HINTS.some((h) => lower.includes(h));
641
+ }
601
642
  async function chatCompletionOpenAI(config, messages, maxTokens, temperature, timeoutMs) {
602
643
  const url = `${config.baseUrl}/chat/completions`;
603
644
  const body = {
@@ -606,6 +647,12 @@ async function chatCompletionOpenAI(config, messages, maxTokens, temperature, ti
606
647
  temperature,
607
648
  max_completion_tokens: maxTokens,
608
649
  };
650
+ // 3.3.12-rc.6: hint the provider to return strict JSON. Critical for
651
+ // z.ai/GLM (silent-empty without it). See JSON_OBJECT_PROVIDER_HINTS
652
+ // doc above.
653
+ if (supportsJsonObjectResponseFormat(config.baseUrl)) {
654
+ body.response_format = { type: 'json_object' };
655
+ }
609
656
  try {
610
657
  const res = await fetch(url, {
611
658
  method: 'POST',
@@ -621,7 +668,28 @@ async function chatCompletionOpenAI(config, messages, maxTokens, temperature, ti
621
668
  throw new Error(`LLM API ${res.status}: ${text.slice(0, 200)}`);
622
669
  }
623
670
  const json = (await res.json());
624
- return json.choices?.[0]?.message?.content ?? null;
671
+ const content = json.choices?.[0]?.message?.content ?? null;
672
+ // 3.3.12-rc.6: loud-on-empty. If the provider returned a 200 with
673
+ // empty content, this almost always means a missing response_format
674
+ // hint or a content-filter. Without this log the silent-empty
675
+ // failure mode (Python rc.23 / plugin rc.5) is invisible to ops.
676
+ if (content === '' || content === null) {
677
+ // Lazy import to avoid circular dep with the registered logger.
678
+ // Fall back to console.warn if logger unavailable.
679
+ const warn = (msg) => {
680
+ try {
681
+ // eslint-disable-next-line @typescript-eslint/no-explicit-any
682
+ globalThis?.console?.warn?.(msg);
683
+ }
684
+ catch {
685
+ /* noop */
686
+ }
687
+ };
688
+ warn(`[totalreclaw][llm-client] provider=${config.baseUrl} model=${config.model} ` +
689
+ `returned empty content (status=200). ` +
690
+ `If using z.ai/GLM/OpenAI-compat, check response_format hint is being sent.`);
691
+ }
692
+ return content;
625
693
  }
626
694
  catch (err) {
627
695
  const msg = err instanceof Error ? err.message : String(err);
package/dist/tr-cli.js CHANGED
@@ -51,7 +51,7 @@ const STATE_PATH = CONFIG.onboardingStatePath;
51
51
  // Auto-synced by skill/scripts/sync-version.mjs from skill/plugin/package.json::version.
52
52
  // Do not edit by hand — running tests will catch drift but the publish workflow
53
53
  // rewrites this constant at the start of every npm/ClawHub publish.
54
- const PLUGIN_VERSION = '3.3.12-rc.4';
54
+ const PLUGIN_VERSION = '3.3.12-rc.6';
55
55
  function die(msg, code = 1) {
56
56
  process.stderr.write(`tr: ${msg}\n`);
57
57
  process.exit(code);
package/import-adapters/types.ts CHANGED
@@ -19,7 +19,7 @@ export interface NormalizedFact {
19
19
  tags?: string[];
20
20
  }
21
21
 
22
- export type ImportSource = 'mem0' | 'mcp-memory' | 'chatgpt' | 'claude' | 'gemini' | 'memoclaw' | 'generic-json' | 'generic-csv';
22
+ export type ImportSource = 'mem0' | 'mcp-memory' | 'chatgpt' | 'claude' | 'gemini';
23
23
 
24
24
  /**
25
25
  * What the user passes to the import tool.
package/index.ts CHANGED
@@ -2255,7 +2255,7 @@ async function handlePluginImportFrom(
2255
2255
  const startTime = Date.now();
2256
2256
 
2257
2257
  const source = params.source as string;
2258
- const validSources = ['mem0', 'mcp-memory', 'chatgpt', 'claude', 'gemini', 'memoclaw', 'generic-json', 'generic-csv'];
2258
+ const validSources = ['mem0', 'mcp-memory', 'chatgpt', 'claude', 'gemini'];
2259
2259
 
2260
2260
  if (!source || !validSources.includes(source)) {
2261
2261
  return { success: false, error: `Invalid source. Must be one of: ${validSources.join(', ')}` };
@@ -2628,7 +2628,7 @@ async function handleBatchImport(
2628
2628
  const offset = (params.offset as number) ?? 0;
2629
2629
  const batchSize = (params.batch_size as number) ?? 25;
2630
2630
 
2631
- const validSources = ['mem0', 'mcp-memory', 'chatgpt', 'claude', 'gemini', 'memoclaw', 'generic-json', 'generic-csv'];
2631
+ const validSources = ['mem0', 'mcp-memory', 'chatgpt', 'claude', 'gemini'];
2632
2632
  if (!source || !validSources.includes(source)) {
2633
2633
  return { success: false, error: `Invalid source. Must be one of: ${validSources.join(', ')}` };
2634
2634
  }
@@ -3156,14 +3156,55 @@ const plugin = {
3156
3156
  // Fix #1 which gates on installs being present).
3157
3157
  const patchResult = patchOpenClawConfig(undefined, pluginVersion ?? undefined);
3158
3158
  if (patchResult === 'patched') {
3159
+ // 3.3.12-rc.6 (auto-QA finding 2026-05-09): previously we only
3160
+ // warned the user to manually restart. That created a silent
3161
+ // hook-failure on the FIRST gateway boot post-install — the
3162
+ // plugin loads with stale in-memory config, hook handlers
3163
+ // never register, auto-extraction never fires, and only a
3164
+ // second manual restart fixes it. First-time users hit this
3165
+ // every install. Auto-extraction QA reproduced it as 2/5
3166
+ // turns missed (hook silently no-op'd on turns 1-3).
3167
+ //
3168
+ // Fix: when the patch wrote anything, fire SIGUSR1 to our own
3169
+ // PID. The gateway accepts SIGUSR1 iff `commands.restart=true`
3170
+ // (the default); see upstream `setGatewaySigusr1RestartPolicy`.
3171
+ // The signal triggers an in-process restart that re-reads the
3172
+ // freshly-patched openclaw.json and registers hook handlers
3173
+ // with `allowConversationAccess=true` honoured.
3174
+ //
3175
+ // Idempotency: second boot reads the patched config and
3176
+ // returns `'unchanged'` from patchOpenClawConfig, so the
3177
+ // signal fires AT MOST once per config-key-change. No restart
3178
+ // loop possible.
3179
+ //
3180
+ // Defer via setImmediate so register() finishes (logger flush
3181
+ // + plugin load record writeback) before the signal lands.
3182
+ // A 250ms setTimeout adds slack for slow disk on Telegram VPS
3183
+ // (Hetzner small VPS tail-latency observed ~120ms on writes).
3184
+ //
3185
+ // Phrase-safety: process.kill on own PID is local-only; no
3186
+ // outbound markers. Already used by `/totalreclaw-restart`
3187
+ // (registered ~400 lines below) under the same scanner-safe
3188
+ // pattern.
3159
3189
  api.logger.warn(
3160
3190
  'TotalReclaw: updated openclaw.json with required 2026.5.x keys ' +
3161
3191
  '(plugins.slots.memory + hooks.allowConversationAccess + ' +
3162
3192
  'channels.telegram.streaming.mode + plugins.bundledDiscovery + ' +
3163
3193
  'plugins.allow + plugins.installs.totalreclaw self-heal). ' +
3164
- 'Gateway restart required for the changes to take effect. ' +
3165
- 'Run `/totalreclaw-restart` or restart the gateway manually.',
3194
+ 'Auto-restarting gateway via SIGUSR1 to apply.',
3166
3195
  );
3196
+ setTimeout(() => {
3197
+ try {
3198
+ process.kill(process.pid, 'SIGUSR1');
3199
+ } catch (err) {
3200
+ const msg = err instanceof Error ? err.message : String(err);
3201
+ api.logger.warn(
3202
+ `TotalReclaw: auto-restart SIGUSR1 emit failed (${msg}). ` +
3203
+ 'Run `/totalreclaw-restart` or restart the gateway manually ' +
3204
+ 'for the patched config to take effect.',
3205
+ );
3206
+ }
3207
+ }, 250);
3167
3208
  } else if (patchResult === 'error') {
3168
3209
  api.logger.warn(
3169
3210
  'TotalReclaw: failed to auto-patch openclaw.json for OpenClaw 2026.5.x ' +
@@ -5296,7 +5337,7 @@ const plugin = {
5296
5337
  properties: {
5297
5338
  source: {
5298
5339
  type: 'string',
5299
- enum: ['mem0', 'mcp-memory', 'chatgpt', 'claude', 'gemini', 'memoclaw', 'generic-json', 'generic-csv'],
5340
+ enum: ['mem0', 'mcp-memory', 'chatgpt', 'claude', 'gemini'],
5300
5341
  description: 'The source system to import from (gemini: Google Takeout HTML; chatgpt: conversations.json or memory text; claude: memory text)',
5301
5342
  },
5302
5343
  api_key: {
package/llm-client.ts CHANGED
@@ -801,6 +801,48 @@ export function isRetryable(errorMessage: string): boolean {
801
801
  // OpenAI-compatible chat completion
802
802
  // ---------------------------------------------------------------------------
803
803
 
804
+ /**
805
+ * Provider base-URL hints for OpenAI-compatible endpoints that honour the
806
+ * `response_format: {"type": "json_object"}` body field. Sending the field
807
+ * to a provider that supports it makes JSON output deterministic; sending
808
+ * it to a provider that does NOT recognise the field is a 400.
809
+ *
810
+ * Why this exists (3.3.12-rc.6, 2026-05-09):
811
+ * z.ai's GLM family (4.5-flash, 5-turbo, 5.1) silently returns EMPTY
812
+ * `message.content` for the merged-extraction prompt unless this hint
813
+ * is set. No error, no warning — the LLM just emits "" instead of the
814
+ * expected `{"topics": [], "facts": []}` JSON. Plugin's parse step
815
+ * then logs `0 raw facts` from a successful-but-empty branch.
816
+ *
817
+ * This bug was found and fixed on the Python (Hermes) side in
818
+ * 2.3.1-rc.23 (see `python/src/totalreclaw/agent/llm_client.py`
819
+ * `_supports_json_object_response_format`) but the plugin TS port did
820
+ * not carry the fix — observed in plugin 3.3.12-rc.5 auto-QA on
821
+ * 2026-05-09: hook + poller both fired correctly but extraction
822
+ * returned 0 facts on every batch despite trajectories containing
823
+ * explicit "I prefer X" / "I work at Y" statements.
824
+ *
825
+ * Mirror of Python's `_supports_json_object_response_format`. Match by
826
+ * substring on a lowercased baseUrl so cosmetic prefix differences
827
+ * (https://, /v1, etc.) don't matter.
828
+ */
829
+ const JSON_OBJECT_PROVIDER_HINTS = [
830
+ 'z.ai',
831
+ 'api.openai.com',
832
+ 'groq.com',
833
+ 'openrouter.ai',
834
+ 'deepseek.com',
835
+ 'mistral.ai',
836
+ 'x.ai',
837
+ 'together.xyz',
838
+ ] as const;
839
+
840
+ export function supportsJsonObjectResponseFormat(baseUrl: string | undefined): boolean {
841
+ if (!baseUrl) return false;
842
+ const lower = baseUrl.toLowerCase();
843
+ return JSON_OBJECT_PROVIDER_HINTS.some((h) => lower.includes(h));
844
+ }
845
+
804
846
  async function chatCompletionOpenAI(
805
847
  config: LLMClientConfig,
806
848
  messages: ChatMessage[],
@@ -817,6 +859,13 @@ async function chatCompletionOpenAI(
817
859
  max_completion_tokens: maxTokens,
818
860
  };
819
861
 
862
+ // 3.3.12-rc.6: hint the provider to return strict JSON. Critical for
863
+ // z.ai/GLM (silent-empty without it). See JSON_OBJECT_PROVIDER_HINTS
864
+ // doc above.
865
+ if (supportsJsonObjectResponseFormat(config.baseUrl)) {
866
+ body.response_format = { type: 'json_object' };
867
+ }
868
+
820
869
  try {
821
870
  const res = await fetch(url, {
822
871
  method: 'POST',
@@ -834,7 +883,31 @@ async function chatCompletionOpenAI(
834
883
  }
835
884
 
836
885
  const json = (await res.json()) as ChatCompletionResponse;
837
- return json.choices?.[0]?.message?.content ?? null;
886
+ const content = json.choices?.[0]?.message?.content ?? null;
887
+
888
+ // 3.3.12-rc.6: loud-on-empty. If the provider returned a 200 with
889
+ // empty content, this almost always means a missing response_format
890
+ // hint or a content-filter. Without this log the silent-empty
891
+ // failure mode (Python rc.23 / plugin rc.5) is invisible to ops.
892
+ if (content === '' || content === null) {
893
+ // Lazy import to avoid circular dep with the registered logger.
894
+ // Fall back to console.warn if logger unavailable.
895
+ const warn = (msg: string) => {
896
+ try {
897
+ // eslint-disable-next-line @typescript-eslint/no-explicit-any
898
+ (globalThis as any)?.console?.warn?.(msg);
899
+ } catch {
900
+ /* noop */
901
+ }
902
+ };
903
+ warn(
904
+ `[totalreclaw][llm-client] provider=${config.baseUrl} model=${config.model} ` +
905
+ `returned empty content (status=200). ` +
906
+ `If using z.ai/GLM/OpenAI-compat, check response_format hint is being sent.`,
907
+ );
908
+ }
909
+
910
+ return content;
838
911
  } catch (err) {
839
912
  const msg = err instanceof Error ? err.message : String(err);
840
913
  throw new Error(`LLM call failed: ${msg}`);
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@totalreclaw/totalreclaw",
3
- "version": "3.3.12-rc.4",
3
+ "version": "3.3.12-rc.6",
4
4
  "description": "End-to-end encrypted, agent-portable memory for OpenClaw and any LLM-agent runtime. XChaCha20-Poly1305 with protobuf v4 + on-chain Memory Taxonomy v1 (claim / preference / directive / commitment / episode / summary).",
5
5
  "type": "module",
6
6
  "keywords": [
@@ -67,7 +67,7 @@
67
67
  "scripts": {
68
68
  "build": "rm -rf dist && tsc -p tsconfig.json --noCheck",
69
69
  "verify-tarball": "node ../scripts/verify-tarball.mjs",
70
- "test": "npx tsx manifest-shape.test.ts && npx tsx config-schema.test.ts && npx tsx config.test.ts && npx tsx relay-headers.test.ts && npx tsx scope-address-visible.test.ts && npx tsx llm-profile-reader.test.ts && npx tsx llm-client.test.ts && npx tsx llm-client-retry.test.ts && npx tsx gateway-url.test.ts && npx tsx retype-setscope.test.ts && npx tsx tool-gating.test.ts && npx tsx onboarding-noninteractive.test.ts && npx tsx pair-cli-json.test.ts && npx tsx pair-qr.test.ts && npx tsx pair-remote-client.test.ts && npx tsx qa-bug-report.test.ts && npx tsx nonce-serialization.test.ts && npx tsx phrase-safety-registry.test.ts && npx tsx test_issue_92_onnx_download_ux.test.ts && npx tsx onboard-pair-only.test.ts && npx tsx import-time-smoke.test.ts && npx tsx install-staging-cleanup.test.ts && npx tsx partial-install-detection.test.ts && npx tsx install-reload-idempotency.test.ts && npx tsx json-stdout-cleanliness.test.ts && npx tsx load-manifest.test.ts && npx tsx url-binding.test.ts && npx tsx fs-helpers.test.ts && npx tsx pair-cli-default-mode.test.ts && npx tsx embedding-fallback-tag.test.ts && npx tsx staging-banner-gate.test.ts && npx tsx restart-auth.test.ts && npx tsx inbound-user-tracker.test.ts && npx tsx register-command-name.test.ts && npx tsx skill-md-hybrid-primary.test.ts && npx tsx tr-cli-json-output.test.ts",
70
+ "test": "npx tsx manifest-shape.test.ts && npx tsx config-schema.test.ts && npx tsx config.test.ts && npx tsx relay-headers.test.ts && npx tsx scope-address-visible.test.ts && npx tsx llm-profile-reader.test.ts && npx tsx llm-client.test.ts && npx tsx llm-client-retry.test.ts && npx tsx llm-client-json-mode.test.ts && npx tsx gateway-url.test.ts && npx tsx retype-setscope.test.ts && npx tsx tool-gating.test.ts && npx tsx onboarding-noninteractive.test.ts && npx tsx pair-cli-json.test.ts && npx tsx pair-qr.test.ts && npx tsx pair-remote-client.test.ts && npx tsx qa-bug-report.test.ts && npx tsx nonce-serialization.test.ts && npx tsx phrase-safety-registry.test.ts && npx tsx test_issue_92_onnx_download_ux.test.ts && npx tsx onboard-pair-only.test.ts && npx tsx import-time-smoke.test.ts && npx tsx install-staging-cleanup.test.ts && npx tsx partial-install-detection.test.ts && npx tsx install-reload-idempotency.test.ts && npx tsx json-stdout-cleanliness.test.ts && npx tsx load-manifest.test.ts && npx tsx url-binding.test.ts && npx tsx fs-helpers.test.ts && npx tsx pair-cli-default-mode.test.ts && npx tsx embedding-fallback-tag.test.ts && npx tsx staging-banner-gate.test.ts && npx tsx restart-auth.test.ts && npx tsx inbound-user-tracker.test.ts && npx tsx register-command-name.test.ts && npx tsx skill-md-hybrid-primary.test.ts && npx tsx tr-cli-json-output.test.ts",
71
71
  "smoke:dist": "npx tsx dist-esm-smoke.test.ts",
72
72
  "check-scanner": "node ../scripts/check-scanner.mjs",
73
73
  "check-version-drift": "node ../scripts/check-version-drift.mjs",
package/skill.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "totalreclaw",
3
- "version": "3.3.12-rc.4",
3
+ "version": "3.3.12-rc.6",
4
4
  "description": "End-to-end encrypted memory for AI agents — portable, yours forever. XChaCha20-Poly1305 E2EE: server never sees plaintext.",
5
5
  "author": "TotalReclaw Team",
6
6
  "license": "MIT",
package/tr-cli.ts CHANGED
@@ -72,7 +72,7 @@ const STATE_PATH = CONFIG.onboardingStatePath;
72
72
  // Auto-synced by skill/scripts/sync-version.mjs from skill/plugin/package.json::version.
73
73
  // Do not edit by hand — running tests will catch drift but the publish workflow
74
74
  // rewrites this constant at the start of every npm/ClawHub publish.
75
- const PLUGIN_VERSION = '3.3.12-rc.4';
75
+ const PLUGIN_VERSION = '3.3.12-rc.6';
76
76
 
77
77
  function die(msg: string, code = 1): never {
78
78
  process.stderr.write(`tr: ${msg}\n`);