dutygate 0.1.0__tar.gz

dutygate-0.1.0/.gitignore

@@ -0,0 +1,19 @@
+__pycache__/
+*.py[cod]
+*.egg-info/
+.venv/
+dist/
+build/
+.coverage
+coverage.xml
+htmlcov/
+.pytest_cache/
+.mypy_cache/
+.ruff_cache/
+.hypothesis/
+node_modules/
+clients/typescript/dist/
+.env
+.env.*
+*.log
+.DS_Store

dutygate-0.1.0/CHANGELOG.md

@@ -0,0 +1,44 @@
+# Changelog
+
+All notable changes are documented here. The format follows
+[Keep a Changelog](https://keepachangelog.com/en/1.1.0/), and the project uses
+[Semantic Versioning](https://semver.org/).
+
+## [Unreleased]
+
+## [0.1.0] - 2026-09-28
+
+### Added
+- Policy pack format (`schema_version: 1`) with validation that reports every error at once,
+  plus warnings.
+- Engine with one backend call per message; `match: any|all`; per-rule thresholds; confident
+  and grey flags; fail-safe error decisions with machine-readable codes.
+- Backends:
+  - TypeSafe Jev: total time budget, retries on 429/529/5xx
+  - replay: fixtures bound to a questions digest
+  - keyword: evaluation baseline only
+- Redaction before any backend call: emails, and card numbers only when they pass a Luhn check.
+- `Gate` facade (sync and async).
+- CLI: `validate`, `run`, `eval` and `serve`.
+- Evaluation harness: metrics per category and per tag, `--save-answers`, `--replay`,
+  threshold `--sweep`, keyword baseline comparison, and `--fail-under` gates for CI.
+- HTTP sidecar:
+  - bearer auth with key rotation, multi-pack routing, a body size limit
+  - Prometheus metrics and an opt-in audit log
+  - JSON logs without message text
+  - Docker image
+- TypeScript client `dutygate-client`: zero dependencies, fail-safe `review` fallback,
+  ESM and CJS builds.
+- `GatedHandler` for any web framework, the LangChain `with_legal_gate` adapter, and LangGraph
+  `gate_node` / `outbound_node` (extra `langgraph`).
+- Packs: `legal-triggers` (inbound) and `outbound-claims` (the bot's replies), bundled in the
+  package with their sample datasets and keyword baselines, and loadable by name.
+- `dutygate init <pack>` copies a bundled pack, its dataset and its keywords into your project
+  to customize. `dutygate eval <pack>` and `--backend keyword` default to the bundled files.
+- Conformance suite (55 cases), run through the engine, the sidecar and the TypeScript client.
+- Synthetic evaluation datasets (159 and 66 rows) with keyword baselines.
+
+### Known limitations
+- The TypeSafe Jev integration follows the published API reference but has not yet been
+  verified against the live API.
+- Question wording and thresholds are untuned.

dutygate-0.1.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Taimoor Rashid
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

dutygate-0.1.0/PKG-INFO

@@ -0,0 +1,240 @@
+Metadata-Version: 2.5
+Name: dutygate
+Version: 0.1.0
+Summary: DutyGate: a chatbot-agnostic gate that flags legal and compliance triggers in inbound messages before the bot replies.
+Project-URL: Homepage, https://github.com/Taimoor2500/Dutygate
+Project-URL: Documentation, https://github.com/Taimoor2500/Dutygate#readme
+Project-URL: Issues, https://github.com/Taimoor2500/Dutygate/issues
+Author: Taimoor Rashid
+License-Expression: MIT
+License-File: LICENSE
+Keywords: chatbot,compliance,guardrail,legal,policy,typesafe
+Classifier: Development Status :: 3 - Alpha
+Classifier: Intended Audience :: Developers
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Topic :: Communications :: Chat
+Classifier: Typing :: Typed
+Requires-Python: >=3.10
+Requires-Dist: click>=8.1
+Requires-Dist: httpx>=0.27
+Requires-Dist: pydantic>=2.6
+Requires-Dist: pyyaml>=6
+Provides-Extra: langchain
+Requires-Dist: langchain-core>=0.3; extra == 'langchain'
+Provides-Extra: langgraph
+Requires-Dist: langgraph>=1.0; extra == 'langgraph'
+Provides-Extra: server
+Requires-Dist: fastapi>=0.110; extra == 'server'
+Requires-Dist: prometheus-client>=0.20; extra == 'server'
+Requires-Dist: uvicorn[standard]>=0.29; extra == 'server'
+Description-Content-Type: text/markdown
+
+# DutyGate
+
+**Catch the legal obligations hiding in ordinary support messages before your chatbot replies.**
+
+![DutyGate demo: legal triggers in customer messages are held before the bot replies](https://raw.githubusercontent.com/Taimoor2500/Dutygate/main/docs/assets/dutygate-demo.gif)
+
+"pls stop texting me" is an opt-out. "I want everything you have on me" is a privacy access
+request. "My lawyer will hear about this" is a legal threat. Topic triage files these under
+*billing* or *angry customer*, keyword lists miss the way people actually write, and an LLM bot
+that answers "Done, you're unsubscribed!" when nothing happened creates the liability itself.
+
+DutyGate is a small, chatbot-agnostic gate placed in front of the reply step. It reads every
+inbound message, asks [TypeSafe Jev](https://docs.typesafe.ai) a batch of yes/no questions in
+**one** call, applies versioned rules from a YAML **policy pack**, and tells your code what to do:
+
+```
+inbound message ──► DUTYGATE ──► continue ──► bot replies normally
+                        ├── route  ──► holding reply + case in the right queue
+                        └── review ──► bot may reply; a human takes a look
+```
+
+- **Flags and routes, never acts.** It does not unsubscribe, delete, refund or reply on its
+  own. Your people and workflows do that.
+- **Recall-first.** A miss can cost a statutory deadline; a false alarm costs a reviewer a few
+  seconds. Low-confidence signals go to `review`, not to `continue`.
+- **Fails safe.** If TypeSafe is down, slow, or returns something malformed, you get a `review`
+  decision with an error code, never an exception or a dropped message.
+- **Auditable.** Rules are versioned YAML, testable offline, with a published evaluation
+  harness and a cross-language conformance suite.
+
+> **Not legal advice.** Which categories matter, their deadlines, and how they are routed depend
+> on your jurisdiction and business. Have counsel review the packs and your holding replies
+> before production use.
+
+## Quickstart
+
+Offline, with no API key. The reference packs, their sample datasets and keyword lists ship
+inside the package, so they can be used by name. The `keyword` backend here is the naive
+baseline, which exists only for evaluation and demos.
+
+```console
+pip install 'dutygate[server]'
+
+dutygate validate legal-triggers outbound-claims
+dutygate run legal-triggers --backend keyword --state "pls stop texting me"
+dutygate eval legal-triggers --backend keyword
+```
+
+With TypeSafe Jev, the real backend:
+
+```console
+export TYPESAFE_API_KEY=...        # from typesafe.ai
+export DUTYGATE_JEV_MODEL=jev-1.13.0   # pin the model in production
+dutygate run legal-triggers --state "I want everything you have on me"
+dutygate eval legal-triggers    # scores Jev on the bundled sample dataset
+```
+
+## Make it yours
+
+Copy a reference pack, with its sample dataset and keyword list, into your project and edit
+it. No code is involved: questions, thresholds, queues and new categories are all YAML.
+
+```console
+dutygate init legal-triggers      # writes legal-triggers.yaml, .dataset.jsonl, .keywords.yaml
+# edit legal-triggers.yaml (and give it your own `name`), add rows to the dataset
+dutygate validate legal-triggers.yaml
+dutygate eval legal-triggers.yaml legal-triggers.dataset.jsonl --save-answers answers.jsonl
+dutygate eval legal-triggers.yaml legal-triggers.dataset.jsonl --replay answers.jsonl --sweep
+```
+
+Then point your bot at your file: `Gate.from_pack("legal-triggers.yaml")`. To measure it on
+real traffic, add your own labeled messages to the dataset; see
+[docs/labeling.md](docs/labeling.md). The format is in [SPEC.md](SPEC.md).
+
+## Use it from your bot
+
+### Python
+
+```python
+from dutygate import Gate, default_holding_reply
+
+gate = Gate.from_pack("packs/legal-triggers.yaml")  # TypeSafe Jev, configured from env
+
+decision = await gate.check_async(message, conversation_id=cid, channel="sms")
+if decision.action == "route":
+    await create_case(decision.flags, message)  # your ticketing / queue
+    reply = default_holding_reply(decision)  # neutral; no bot improvisation
+elif decision.action == "review":
+    await create_review_task(decision.flags, message)
+    if decision.error:  # the gate could not judge it
+        await enqueue_rescan(message)
+    reply = await bot.reply(message)
+else:
+    reply = await bot.reply(message)
+```
+
+`GatedHandler` (in `dutygate.adapters.webhook`) wraps exactly this logic for any web
+framework, `with_legal_gate` does the same for LangChain runnables, and
+`gate_node` / `outbound_node` add DutyGate to a LangGraph graph. See
+[docs/adapters.md](docs/adapters.md).
+
+### Any language: the HTTP sidecar
+
+```console
+DUTYGATE_SIDECAR_KEYS=change-me TYPESAFE_API_KEY=... \
+  dutygate serve packs/legal-triggers.yaml packs/outbound-claims.yaml --host 0.0.0.0
+
+curl -H 'Authorization: Bearer change-me' -H 'content-type: application/json' \
+  -d '{"message": "pls stop texting me", "channel": "sms"}' localhost:8080/v1/gate
+```
+
+Or use the Docker image, the [TypeScript client](clients/typescript) (`dutygate-client`), or
+any no-code tool ([n8n / Zapier recipe](examples/n8n/README.md)). See
+[docs/sidecar.md](docs/sidecar.md).
+
+## The decision
+
+```json
+{
+  "id": "dec_4f1c2a9e8b7d4c3fa1e2b3c4d5e6f708",
+  "action": "route",
+  "primary": "privacy_request",
+  "flags": [
+    {"category": "privacy_request", "rule_id": "privacy-request", "queue": "privacy",
+     "priority": "high", "confidence": 0.93, "level": "confident"},
+    {"category": "opt_out", "rule_id": "opt-out", "queue": "compliance",
+     "priority": "high", "confidence": 0.31, "level": "grey"}
+  ],
+  "error": null,
+  "policy": {"name": "legal-triggers", "version": "0.1.0"},
+  "backend": {"name": "typesafe-jev", "model": "jev-1.13.0"},
+  "conversation_id": "c_123",
+  "latency_ms": 212
+}
+```
+
+| `action` | meaning | your bot should |
+|---|---|---|
+| `continue` | no trigger | reply normally |
+| `route` | at least one confident trigger | create a case, send a neutral holding reply, and not improvise about the trigger |
+| `review` | a grey-zone signal, or the gate could not judge (`error` is set) | reply if you like, queue a human review, never auto-close; re-scan later if `error` is set |
+
+`primary` is the first confident flag in the pack's rule order (or the first grey one).
+Error codes and every field are specified in [SPEC.md](SPEC.md).
+
+## What it detects
+
+| pack | categories |
+|---|---|
+| `legal-triggers` (inbound) | `legal_threat`, `regulator_complaint`, `privacy_request`, `chargeback_dispute`, `opt_out`, `accessibility_request` |
+| `outbound-claims` (the bot's own reply) | `admits_liability`, `claims_action_completed`, `legal_statement`, `promises_outcome` |
+
+One message can raise several flags ("erase my data and stop the emails"). Queues and
+priorities are examples: map them to your own systems. Packs are plain YAML, so you can add
+categories or tighten questions without touching code ([SPEC.md](SPEC.md)).
+
+## Surfaces
+
+| surface | status |
+|---|---|
+| Python library (`Gate`, `evaluate`) | ✅ |
+| CLI: `validate`, `run`, `eval`, `serve` | ✅ |
+| HTTP sidecar + Docker image | ✅ |
+| TypeScript client `dutygate-client` | ✅ passes the conformance suite end to end |
+| `GatedHandler` for any web framework (FastAPI, Flask examples) | ✅ |
+| LangChain `with_legal_gate` | ✅ |
+| LangGraph gate and outbound nodes | ✅ |
+
+Every surface must reproduce [`conformance/cases.json`](conformance/cases.json).
+
+## Evaluation
+
+The repo ships synthetic labeled datasets, a deliberately naive keyword baseline, and a
+harness that scores any backend, saves raw answers, and sweeps thresholds offline.
+
+On the bundled `legal-triggers` set, the keyword baseline catches **48%** of triggers
+(0% of non-English ones) and false-alarms on **13%** of clean messages. The datasets were
+written by the same author as the baseline, so they are a sanity check, not a benchmark:
+measure on your own traffic before you rely on this. See [evals/README.md](evals/README.md).
+
+**If Jev does not beat a keyword list on your data, do not pay for it.**
+
+## Status
+
+| item | state |
+|---|---|
+| Packs, engine, validator, CLI, eval harness, sidecar, TS client, adapters | built and tested offline (Python coverage ≥ 90%) |
+| Live TypeSafe Jev integration | built against the published API reference and tested with mocked HTTP; **not yet verified against the live API** (`pytest -m live`) |
+| Question wording and thresholds | untuned; tune with `dutygate eval --save-answers` then `--replay --sweep` |
+
+## Docs
+
+- [SPEC.md](SPEC.md): pack format, evaluation algorithm, decision contract (normative)
+- [docs/integration.md](docs/integration.md): host contract, holding replies, failure behavior, latency patterns
+- [docs/sidecar.md](docs/sidecar.md): HTTP API, auth, metrics, audit log, Docker
+- [docs/adapters.md](docs/adapters.md): `GatedHandler`, LangChain, LangGraph
+- [docs/privacy.md](docs/privacy.md): what leaves your system, redaction, prompt injection
+- [docs/labeling.md](docs/labeling.md): building a real evaluation set
+- [CONTRIBUTING.md](CONTRIBUTING.md) · [SECURITY.md](SECURITY.md) · [CHANGELOG.md](CHANGELOG.md)
+
+## License
+
+MIT. See [LICENSE](LICENSE).

dutygate-0.1.0/README.md

@@ -0,0 +1,203 @@
+# DutyGate
+
+**Catch the legal obligations hiding in ordinary support messages before your chatbot replies.**
+
+![DutyGate demo: legal triggers in customer messages are held before the bot replies](https://raw.githubusercontent.com/Taimoor2500/Dutygate/main/docs/assets/dutygate-demo.gif)
+
+"pls stop texting me" is an opt-out. "I want everything you have on me" is a privacy access
+request. "My lawyer will hear about this" is a legal threat. Topic triage files these under
+*billing* or *angry customer*, keyword lists miss the way people actually write, and an LLM bot
+that answers "Done, you're unsubscribed!" when nothing happened creates the liability itself.
+
+DutyGate is a small, chatbot-agnostic gate placed in front of the reply step. It reads every
+inbound message, asks [TypeSafe Jev](https://docs.typesafe.ai) a batch of yes/no questions in
+**one** call, applies versioned rules from a YAML **policy pack**, and tells your code what to do:
+
+```
+inbound message ──► DUTYGATE ──► continue ──► bot replies normally
+                        ├── route  ──► holding reply + case in the right queue
+                        └── review ──► bot may reply; a human takes a look
+```
+
+- **Flags and routes, never acts.** It does not unsubscribe, delete, refund or reply on its
+  own. Your people and workflows do that.
+- **Recall-first.** A miss can cost a statutory deadline; a false alarm costs a reviewer a few
+  seconds. Low-confidence signals go to `review`, not to `continue`.
+- **Fails safe.** If TypeSafe is down, slow, or returns something malformed, you get a `review`
+  decision with an error code, never an exception or a dropped message.
+- **Auditable.** Rules are versioned YAML, testable offline, with a published evaluation
+  harness and a cross-language conformance suite.
+
+> **Not legal advice.** Which categories matter, their deadlines, and how they are routed depend
+> on your jurisdiction and business. Have counsel review the packs and your holding replies
+> before production use.
+
+## Quickstart
+
+Offline, with no API key. The reference packs, their sample datasets and keyword lists ship
+inside the package, so they can be used by name. The `keyword` backend here is the naive
+baseline, which exists only for evaluation and demos.
+
+```console
+pip install 'dutygate[server]'
+
+dutygate validate legal-triggers outbound-claims
+dutygate run legal-triggers --backend keyword --state "pls stop texting me"
+dutygate eval legal-triggers --backend keyword
+```
+
+With TypeSafe Jev, the real backend:
+
+```console
+export TYPESAFE_API_KEY=...        # from typesafe.ai
+export DUTYGATE_JEV_MODEL=jev-1.13.0   # pin the model in production
+dutygate run legal-triggers --state "I want everything you have on me"
+dutygate eval legal-triggers    # scores Jev on the bundled sample dataset
+```
+
+## Make it yours
+
+Copy a reference pack, with its sample dataset and keyword list, into your project and edit
+it. No code is involved: questions, thresholds, queues and new categories are all YAML.
+
+```console
+dutygate init legal-triggers      # writes legal-triggers.yaml, .dataset.jsonl, .keywords.yaml
+# edit legal-triggers.yaml (and give it your own `name`), add rows to the dataset
+dutygate validate legal-triggers.yaml
+dutygate eval legal-triggers.yaml legal-triggers.dataset.jsonl --save-answers answers.jsonl
+dutygate eval legal-triggers.yaml legal-triggers.dataset.jsonl --replay answers.jsonl --sweep
+```
+
+Then point your bot at your file: `Gate.from_pack("legal-triggers.yaml")`. To measure it on
+real traffic, add your own labeled messages to the dataset; see
+[docs/labeling.md](docs/labeling.md). The format is in [SPEC.md](SPEC.md).
+
+## Use it from your bot
+
+### Python
+
+```python
+from dutygate import Gate, default_holding_reply
+
+gate = Gate.from_pack("packs/legal-triggers.yaml")  # TypeSafe Jev, configured from env
+
+decision = await gate.check_async(message, conversation_id=cid, channel="sms")
+if decision.action == "route":
+    await create_case(decision.flags, message)  # your ticketing / queue
+    reply = default_holding_reply(decision)  # neutral; no bot improvisation
+elif decision.action == "review":
+    await create_review_task(decision.flags, message)
+    if decision.error:  # the gate could not judge it
+        await enqueue_rescan(message)
+    reply = await bot.reply(message)
+else:
+    reply = await bot.reply(message)
+```
+
+`GatedHandler` (in `dutygate.adapters.webhook`) wraps exactly this logic for any web
+framework, `with_legal_gate` does the same for LangChain runnables, and
+`gate_node` / `outbound_node` add DutyGate to a LangGraph graph. See
+[docs/adapters.md](docs/adapters.md).
+
+### Any language: the HTTP sidecar
+
+```console
+DUTYGATE_SIDECAR_KEYS=change-me TYPESAFE_API_KEY=... \
+  dutygate serve packs/legal-triggers.yaml packs/outbound-claims.yaml --host 0.0.0.0
+
+curl -H 'Authorization: Bearer change-me' -H 'content-type: application/json' \
+  -d '{"message": "pls stop texting me", "channel": "sms"}' localhost:8080/v1/gate
+```
+
+Or use the Docker image, the [TypeScript client](clients/typescript) (`dutygate-client`), or
+any no-code tool ([n8n / Zapier recipe](examples/n8n/README.md)). See
+[docs/sidecar.md](docs/sidecar.md).
+
+## The decision
+
+```json
+{
+  "id": "dec_4f1c2a9e8b7d4c3fa1e2b3c4d5e6f708",
+  "action": "route",
+  "primary": "privacy_request",
+  "flags": [
+    {"category": "privacy_request", "rule_id": "privacy-request", "queue": "privacy",
+     "priority": "high", "confidence": 0.93, "level": "confident"},
+    {"category": "opt_out", "rule_id": "opt-out", "queue": "compliance",
+     "priority": "high", "confidence": 0.31, "level": "grey"}
+  ],
+  "error": null,
+  "policy": {"name": "legal-triggers", "version": "0.1.0"},
+  "backend": {"name": "typesafe-jev", "model": "jev-1.13.0"},
+  "conversation_id": "c_123",
+  "latency_ms": 212
+}
+```
+
+| `action` | meaning | your bot should |
+|---|---|---|
+| `continue` | no trigger | reply normally |
+| `route` | at least one confident trigger | create a case, send a neutral holding reply, and not improvise about the trigger |
+| `review` | a grey-zone signal, or the gate could not judge (`error` is set) | reply if you like, queue a human review, never auto-close; re-scan later if `error` is set |
+
+`primary` is the first confident flag in the pack's rule order (or the first grey one).
+Error codes and every field are specified in [SPEC.md](SPEC.md).
+
+## What it detects
+
+| pack | categories |
+|---|---|
+| `legal-triggers` (inbound) | `legal_threat`, `regulator_complaint`, `privacy_request`, `chargeback_dispute`, `opt_out`, `accessibility_request` |
+| `outbound-claims` (the bot's own reply) | `admits_liability`, `claims_action_completed`, `legal_statement`, `promises_outcome` |
+
+One message can raise several flags ("erase my data and stop the emails"). Queues and
+priorities are examples: map them to your own systems. Packs are plain YAML, so you can add
+categories or tighten questions without touching code ([SPEC.md](SPEC.md)).
+
+## Surfaces
+
+| surface | status |
+|---|---|
+| Python library (`Gate`, `evaluate`) | ✅ |
+| CLI: `validate`, `run`, `eval`, `serve` | ✅ |
+| HTTP sidecar + Docker image | ✅ |
+| TypeScript client `dutygate-client` | ✅ passes the conformance suite end to end |
+| `GatedHandler` for any web framework (FastAPI, Flask examples) | ✅ |
+| LangChain `with_legal_gate` | ✅ |
+| LangGraph gate and outbound nodes | ✅ |
+
+Every surface must reproduce [`conformance/cases.json`](conformance/cases.json).
+
+## Evaluation
+
+The repo ships synthetic labeled datasets, a deliberately naive keyword baseline, and a
+harness that scores any backend, saves raw answers, and sweeps thresholds offline.
+
+On the bundled `legal-triggers` set, the keyword baseline catches **48%** of triggers
+(0% of non-English ones) and false-alarms on **13%** of clean messages. The datasets were
+written by the same author as the baseline, so they are a sanity check, not a benchmark:
+measure on your own traffic before you rely on this. See [evals/README.md](evals/README.md).
+
+**If Jev does not beat a keyword list on your data, do not pay for it.**
+
+## Status
+
+| item | state |
+|---|---|
+| Packs, engine, validator, CLI, eval harness, sidecar, TS client, adapters | built and tested offline (Python coverage ≥ 90%) |
+| Live TypeSafe Jev integration | built against the published API reference and tested with mocked HTTP; **not yet verified against the live API** (`pytest -m live`) |
+| Question wording and thresholds | untuned; tune with `dutygate eval --save-answers` then `--replay --sweep` |
+
+## Docs
+
+- [SPEC.md](SPEC.md): pack format, evaluation algorithm, decision contract (normative)
+- [docs/integration.md](docs/integration.md): host contract, holding replies, failure behavior, latency patterns
+- [docs/sidecar.md](docs/sidecar.md): HTTP API, auth, metrics, audit log, Docker
+- [docs/adapters.md](docs/adapters.md): `GatedHandler`, LangChain, LangGraph
+- [docs/privacy.md](docs/privacy.md): what leaves your system, redaction, prompt injection
+- [docs/labeling.md](docs/labeling.md): building a real evaluation set
+- [CONTRIBUTING.md](CONTRIBUTING.md) · [SECURITY.md](SECURITY.md) · [CHANGELOG.md](CHANGELOG.md)
+
+## License
+
+MIT. See [LICENSE](LICENSE).