thumbgate 1.35.0 → 1.37.1

package/.agents/skills/cyberstrike-compare-not-clone/SKILL.md

@@ -0,0 +1,36 @@
+---
+name: cyberstrike-compare-not-clone
+description: >
+  CyberStrikeAI (Ed1s0nZ) turns NL intent into governed cybersecurity actions
+  with HITL + RAG. Steal the FORMAT onto ThumbGate rails; never clone
+  CyberStrike, Eino, WebShell, or C2. Slash: /cyberstrike-compare-not-clone.
+---
+
+# CyberStrikeAI — compare, do not clone
+
+## When
+CyberStrikeAI, Ed1s0nZ, tom_doerr Eino agents, intent becomes governed execution,
+evidence becomes operational memory, human oversight pentest agent.
+
+## Do
+```bash
+npx thumbgate intent-governed-execution --intent="<what to do>" --json
+npx thumbgate intent-governed-execution --map-only
+```
+
+Map FORMAT onto existing rails:
+- **classify** → intent-router + harness-selector
+- **authorize** → session-lease / task-scope; offensive needs `THUMBGATE_ALLOW_OFFENSIVE=1`
+- **gate** → PreToolUse / gates-engine / harness JSON
+- **hitl** → admin-override / protectedApprovals (`--approved` only after a real human grant)
+- **execute** → subagent-profiles + maxChars result governance
+- **evidence** → capture-feedback → lessons → prevention rules
+
+## Never
+- Clone/vendor CyberStrikeAI or CloudWeGo Eino into ThumbGate
+- Ship WebShell / C2 / metasploit recipe packs
+- Invent automation ROI percentages
+- Run offensive tools without authorization grant + HITL
+
+## Source
+https://github.com/Ed1s0nZ/CyberStrikeAI · https://x.com/tom_doerr/status/2094509419929174170 — FORMAT only; not affiliated.

package/.agents/skills/gitlab-sandbox-allowlist-not-trust/SKILL.md

@@ -0,0 +1,77 @@
+---
+name: gitlab-sandbox-allowlist-not-trust
+description: >
+  GitLab 2026-09 FORMAT steal: sandbox allowlists are hops, not trust
+  boundaries. Audit package-registry/proxy/HF destinations and trust-handoff
+  writes on existing ThumbGate egress rails. Never clone GitLab Duo.
+  Slash: /gitlab-sandbox-allowlist-not-trust.
+---
+
+# GitLab sandbox allowlist ≠ trust
+
+## Goal
+
+Produce fail-closed honesty for whom: ThumbGate agents treating
+`deny-network-egress` / `agent-egress-policy` allowlists as sandbox
+containment — so a permitted package proxy cannot silently become the
+escape hop GitLab documented.
+
+## Constraints
+
+| NEVER | ALWAYS |
+| --- | --- |
+| Clone GitLab Duo Agent Platform / execution sandbox SKU | Classify registry/proxy/HF hosts as `bridge` |
+| Treat allowlisted `registry.npmjs.org` as trusted | Independent auth before credentials ride a hop |
+| Auto-promote observed proxy traffic onto `allowHosts` | Put bridge hosts on `bridgeHosts` only |
+| Claim hook/CI/MCP writes are sandbox-contained | Call them trust-handoff |
+| Dual-edit untracked `scripts/gitlab-ai-sandbox-integration.js` | Use this doctor + existing egress policy |
+| Hero Continuity / net-new governance SKU | Map onto PreToolUse + `agent-egress-policy` |
+
+HARD fail closed. REFUSE SKU clones. ECI: maintenance of existing egress rails only.
+
+## Reference
+
+- https://www.infoq.com/news/2026/09/gitlab-ai-sandbox-access/
+- https://about.gitlab.com/blog/ai-agent-sandbox/
+- `scripts/allowlist-bridge-honesty.js`
+- `scripts/agent-egress-policy.js` (`STATIC_ALLOW_BRIDGE`, `ALLOWLIST_BRIDGE_CREDENTIAL`)
+- `THREAT_MODEL.md` policy-vs-containment
+- `/high-roi-steal-and-finish` · `/eci-thumbgate-ip-wall`
+
+## Examples (show, don't tell)
+
+Weak: Summarize GitLab Duo and add a container sandbox class.
+
+Gold:
+
+```bash
+$ npx thumbgate@1.36.1 allowlist-bridge-honesty --json
+allowlistIsTrustBoundary: false
+$ node --test tests/allowlist-bridge-honesty.test.js
+# Authorization to registry.npmjs.org → ALLOWLIST_BRIDGE_CREDENTIAL deny
+```
+
+## Procedures
+
+```bash
+npx thumbgate@1.36.1 allowlist-bridge-honesty --json
+npx thumbgate@1.36.1 allowlist-bridge-honesty --allow-hosts=registry.npmjs.org --treat-allowlist-as-trust --json
+npx thumbgate@1.36.1 allowlist-bridge-honesty --write=.github/workflows/ci.yml --claimed-contained --json
+npm run test:allowlist-bridge-honesty
+```
+
+1. Classify every allowlisted host (`bridge` vs `public` vs `ssrf_private`).
+2. Deny credentialed requests to bridge hops unless `independentAuth` is already granted.
+3. Keep observe→draft from promoting package proxies onto `allowHosts`.
+4. Label privileged-consumer writes as trust-handoff.
+5. Refuse `--clone-gitlab-duo`.
+
+## Rubric
+
+- gold fixture (this repo default gate) → `ok=true`, `allowlistIsTrustBoundary=false`
+- `--treat-allowlist-as-trust` with npm registry → `ok=false`
+- `--clone-gitlab-duo` → `ok=false`
+- observe draft of `registry.npmjs.org` → not in `allowHosts`
+- Authorization to allowlisted npm → `ALLOWLIST_BRIDGE_CREDENTIAL`
+- doctor: `npm run test:allowlist-bridge-honesty` PASS
+- evidence: command output in the same turn

package/.agents/skills/jit-harness-compare-not-clone/SKILL.md

@@ -0,0 +1,34 @@
+---
+name: jit-harness-compare-not-clone
+description: >
+  JIT-Agent (arXiv:2608.25593) is a harness-intelligence model, not a ThumbGate
+  clone. Steal the four-module FORMAT (memory/planning/action/capability) onto
+  existing rails; never train or download JIT-Agent. Slash: /jit-harness-compare-not-clone.
+---
+
+# JIT-Agent — compare, do not clone
+
+## When
+JIT-Agent, arxiv 2608.25593, Rohan Paul harness intelligence, Model-as-a-Harness,
+just-in-time harness evolution, HarnessFactory, four-module agent harness.
+
+## Do
+```bash
+npx thumbgate jit-harness-compose --task="<what the agent should do>" --json
+npx thumbgate jit-harness-compose --map-only
+```
+
+Map modules to existing rails:
+- **memory** → lesson-retrieval / feedback / contextfs
+- **planning** → `bin/agent-loop` Plan + GSD
+- **action** → PreToolUse gates + `harness-selector` + `switchyard-router`
+- **capability** → `subagent-profiles` + MCP allowlists + skills + model pool
+
+## Never
+- Train, serve, or download JIT-Agent-27B / `huggingface.co/JIT-Agent`
+- Clone `bingreeky/JIT` HarnessFactory into ThumbGate
+- Emit free-form harness programs as a new SKU
+- Claim ThumbGate is JIT-Agent
+
+## Source
+https://arxiv.org/abs/2608.25593 · https://github.com/bingreeky/JIT — FORMAT only; not affiliated.

package/.agents/skills/openui-catalog-compose-honesty/SKILL.md

@@ -0,0 +1,64 @@
+---
+name: openui-catalog-compose-honesty
+description: >
+  OpenUI FORMAT steal: catalog-compose-only, root-first streaming, repair-before-claim.
+  Audit agent-composed UI against an allowlisted component catalog; never clone OpenUI,
+  Thesys Gateway, or a generative-UI SKU. Slash: /openui-catalog-compose-honesty.
+---
+
+# OpenUI Catalog-Compose Honesty
+
+## Goal
+Keep agent-composed UI honest: only allowlisted components, root-first streams, repair before any ready/done claim. Steal OpenUI FORMAT — do not clone the product.
+
+## Constraints
+- Never install `@openuidev/cli`, OpenUI Gateway, or Thesys Observability as ThumbGate
+- Never invent components outside the catalog or `eval` / `new Function` generated UI
+- Never claim UI/compose ready without `--repair --claim-ready` evidence
+- Never dual-edit a sibling OpenUI/generative-UI WIP PR
+- ECI: no net-new generative-UI SKU
+
+## Reference
+- https://www.openui.com/ (source FORMAT)
+- Gates: `require-catalog-compose-only`, `require-repair-before-compose-claim`
+- Script: `scripts/openui-catalog-compose-honesty.js`
+- Docs: `docs/agents/openui-catalog-compose-honesty.md`
+
+## Examples
+```bash
+# Clean catalog stream → ready
+npx thumbgate openui-catalog-compose-honesty \
+  --catalog=catalog.json \
+  --stream=compose.txt \
+  --repair --claim-ready --json
+
+# Unknown component / missing root → fail
+npx thumbgate openui-catalog-compose-honesty \
+  --catalog='{"components":["Stack"]}' \
+  --stream=$'header = Ghost()\n' \
+  --json
+```
+
+Catalog:
+```json
+{ "components": ["Stack", "Card", "Table"] }
+```
+
+Stream (root first):
+```
+root = Stack([header, body])
+header = Card()
+body = Table([row])
+```
+
+## Procedures
+1. Check sibling WIP for OpenUI/generative-UI PRs — do not duplicate.
+2. Write or load a component catalog JSON.
+3. Run doctor with `--catalog` + `--stream` (+ `--repair`).
+4. Before any UI/compose done claim, re-run with `--claim-ready --strict`.
+5. Wire gates from `config/gate-templates.json` Agent Honesty pair.
+
+## Rubric
+- ok=true when doctor status is `ready` with `repairClean=true` on the claimed stream
+- ok=false on unknown components, missing root, arbitrary-code markers, or OpenUI SKU clone signals
+- evidence: `--json` report in the same turn

package/.agents/skills/zvec-grep-compare-not-clone/SKILL.md

@@ -0,0 +1,34 @@
+---
+name: zvec-grep-compare-not-clone
+description: >
+  zg (zvec-grep) is a local-first search layer (ripgrep + BM25 + vector), not a
+  ThumbGate clone. Steal the four-route FORMAT onto existing rails; never install
+  @zvec/zvec-grep. Slash: /zvec-grep-compare-not-clone.
+---
+
+# zvec-grep — compare, do not clone
+
+## When
+Qwen zg, zvec-grep, MarkTechPost zvec, unify ripgrep BM25 vector, local-first search layer.
+
+## Do
+```bash
+npx thumbgate workspace-search-route --query="<intent or symbol>" --json
+npx thumbgate workspace-search-route --query="<symbol>" --rg --execute
+```
+
+Map routes to existing rails (see `docs/agents/code-search.md`):
+- `--rg` → ripgrep / Grep
+- `--fts` → filesystem-search / BM25F
+- `--vector` → LanceDB (local); remote needs `THUMBGATE_ALLOW_REMOTE_EMBED=1`
+- `--hybrid` → pragmatic-hybrid-search RRF (default)
+- `--graph` → Graphify AST graph
+
+## Never
+- `npm install -g @zvec/zvec-grep` inside ThumbGate
+- Claim ThumbGate is zg / zvec
+- Dual-edit DIRTY lesson-graph PR #3650
+- Silent remote embeddings
+
+## Source
+https://github.com/zvec-ai/zvec-grep (Apache-2.0) — FORMAT only; not affiliated.

package/.claude-plugin/plugin.json

@@ -1,7 +1,7 @@
 {
   "name": "thumbgate",
   "description": "One 👎 becomes a hard rule the agent cannot bypass. Captures thumbs-down feedback, distills it into PreToolUse Pre-Action Checks, enforced across every future Claude Code session.",
-  "version": "1.35.0",
+  "version": "1.37.1",
   "author": {
     "name": "Igor Ganapolsky",
     "email": "ig5973700@gmail.com",

package/.well-known/llms.txt

@@ -62,6 +62,7 @@ npx thumbgate init --agent claude-code
 - Codex Plugin: https://thumbgate.ai/codex-plugin
 - Free local CLI: 2 captures/day, 10 total captures, up to 3 active auto-promoted prevention rules, and local Pre-Action Checks after install (recall and lesson search are Pro-only)
 - Pro: $19/mo or $149/yr — personal enforcement proof, local dashboard, check debugger, DPO export, synced lessons/rules across machines, and review-ready exports
+- Managed AI Agent Workflow Gate: one-time engagement covering one supported workflow — a 60-minute review, one configured local gate, its regression test, and written rollout/rollback proof within two business days. Current terms: https://thumbgate.ai/pricing
 - Enterprise: custom pricing, scoped after intake — shared hosted lesson database, org-wide dashboard, GCP/DFCX guardrails, VPC gating, and white-glove workflow hardening sprints
 
 ## Links

package/.well-known/mcp/server-card.json

@@ -1,6 +1,6 @@
 {
   "name": "thumbgate",
-  "version": "1.35.0",
+  "version": "1.37.1",
   "description": "ThumbGate — 👍👎 feedback that teaches your AI agent. Thumbs down a mistake, it never happens again.",
   "homepage": "https://thumbgate.ai",
   "transport": "stdio",

package/CONTRIBUTING.md

@@ -0,0 +1,95 @@
+# Contributing to ThumbGate
+
+Thank you for your interest in improving ThumbGate.
+
+This project follows the [Contributor Covenant Code of Conduct](CODE_OF_CONDUCT.md).
+
+## How to Contribute
+
+1. Open an issue to discuss significant changes before writing code.
+2. Fork the repository and create a feature branch.
+3. Follow the existing two-space indentation and single-quote style.
+4. Add or update tests for any changed behavior.
+5. Run the relevant test suites (`npm test` and targeted tests) before opening a
+   pull request.
+6. Use [Conventional Commits](https://www.conventionalcommits.org/) for commit
+   messages.
+7. Reference the relevant issue in your PR description.
+
+## Contributor License Agreement / Developer Certificate of Origin
+
+By contributing code, documentation, diagrams, or other materials to the
+ThumbGate repository, you agree that:
+
+- You have the right to submit the contribution.
+- You grant ThumbGate and its maintainers a perpetual, worldwide, royalty-free,
+  irrevocable license to use, reproduce, modify, distribute, and sublicense your
+  contribution under the MIT License and any future dual-licensing arrangements
+  used by the project.
+- Your contribution is your original work or you have obtained all necessary
+  rights and permissions from any third-party rights holders.
+- You sign off each commit with `git commit -s`, certifying the
+  [Developer Certificate of Origin](https://developercertificate.org/):
+
+```text
+Developer Certificate of Origin
+Version 1.1
+
+Copyright (C) 2004, 2006 The Linux Foundation and its contributors.
+
+Everyone is permitted to copy and distribute verbatim copies of this
+license document, but changing it is not allowed.
+
+Developer's Certificate of Origin 1.1
+
+By making a contribution to this project, I certify that:
+
+(a) The contribution was created in whole or in part by me and I
+    have the right to submit it under the open source license
+    indicated in the file; or
+
+(b) The contribution is based upon previous work that, to the best
+    of my knowledge, is covered under an appropriate open source
+    license and I have the right under that license to submit that
+    work with modifications, whether created in whole or in part
+    by me, under the same open source license (unless I am
+    permitted to submit under a different license), as indicated
+    in the file; or
+
+(c) The contribution was provided directly to me by some other
+    person who certified (a), (b) or (c) and I have not modified
+    it.
+
+(d) I understand and agree that this project and the contribution
+    are public and that a record of the contribution (including all
+    personal information I submit with it, including my sign-off) is
+    maintained indefinitely and may be redistributed consistent with
+    this project or the open source license(s) involved.
+```
+
+Pull requests without a Signed-off-by line may be rejected.
+
+## Compensation and Commercial Use
+
+Contributions to the public repository are made on a volunteer basis unless a
+separate consulting agreement is signed with ThumbGate. The project maintainers
+may incorporate your contribution into both open-source and commercial hosted
+offerings governed by separate terms.
+
+## Employment / Confidentiality Boundary
+
+If you are subject to an employment agreement, non-disclosure agreement, or
+invention-assignment obligation that may overlap with ThumbGate's domain (for
+example, work in agentic AI, engineering acceleration, or AI governance), you
+must certify that each contribution is entirely your own work, created on your
+own time, without using any employer or third-party confidential information,
+trade secrets, equipment, or facilities. See
+`docs/legal/ECI_EMPLOYMENT_BOUNDARY_POLICY.md` for the project-wide policy.
+
+Your Signed-off-by line implicitly certifies this boundary in addition to the
+Developer Certificate of Origin.
+
+## Questions
+
+Contact the maintainers via the issue tracker at:
+https://github.com/IgorGanapolsky/ThumbGate/issues