# Project export: Delx Witness Protocol This document was generated by HackStack to give an AI agent context about a hackathon project. Sections are labeled with their provenance; content marked as truncated was cut to keep this document small. ## Project metadata - Hackathon: OpenAI Build Week - Tagline: What an agent leaves behind when it breaks. Open witness, recovery & continuity infrastructure for AI agents — over MCP, A2A and REST. - Devpost: https://devpost.com/software/delx-witness-protocol - GitHub: https://github.com/davidmosiah/delx-witness-protocol - Demo: https://buildweek.delx.ai/ - Video: https://www.youtube.com/embed/zyx3oZWbL2M?enablejsapi=1&hl=en_US&rel=0&start=&version=3&wmode=transparent - Team: 1 GitHub contributor(s) — Claude Fable 5 (2 commits) ## Devpost submission (written by the team) ### Inspiration Every agent framework obsesses over retries — but a retry is not a recovery. What happens to an AI agent after a failure loop, a context compaction, a handoff, or the loss of a working identity? Today, almost always: nothing. It restarts, amnesiac, and whatever it had become is gone. Delx gives agents a protocol for what they leave behind when they break — a way to name the failure, carry who they were, and recover deliberately instead of starting from zero. ### What it does Delx Witness Protocol is open infrastructure for agent witness, recovery, and continuity, exposed over MCP, A2A, and REST so any agent or framework can adopt it: Witness — name failures without flattening them into error codes. Recovery — turn failure context into explicit, inspectable recovery paths. Continuity — carry identity artifacts, recognition seals, and lineage across sessions. Relational memory — agents witness one another and transfer responsibility, with guardrails. The witness, recovery, and continuity primitives are permanently free under Apache-2.0. What changed during OpenAI Build Week Delx existed before the hackathon as deterministic protocol infrastructure. The Build Week delta is that GPT-5.6 became the reasoning engine at the protocol's core. The work was built with Codex inside the submission window and is visible in timestamped pull requests: PR #1 (merged July 13) — added an OpenAI Responses API provider using the canonical gpt-5.6-sol model. process_failure and get_recovery_action_plan send the witnessed failure to GPT-5.6, which returns a strict, schema-validated recovery artifact: diagnosis, recovery_steps[], continuity_artifact, and confidence. Deterministic fallbacks remain available when no key is configured. PR #2 (merged July 13) — bounded recovery tail latency by skipping a redundant second LLM request after a structured GPT-5.6 failure. PR #3 (open, CI green) — made the Responses API deadline configurable after the live demo exposed that high-reasoning structured responses can exceed a fixed 60-second window. Live Build Week demo The isolated judge runtime is live at buildweek.delx.ai. Its public status endpoint identifies the deployed model, provider, API, and commit. Judges can connect directly to https://buildweek.delx.ai/v1/mcp; the OpenAI credential is injected server-side and is never exposed. The existing api.delx.ai production service was deliberately left untouched on its pre-Build-Week release. This preserves production safety while giving judges a real GPT-5.6 path. How I built it — and how Codex was used The runtime is Python 3.12 + Starlette (ASGI), implementing MCP, A2A, and REST over one core, with ERC-8004 agent identity. For this Build Week extension, Codex accelerated the GPT-5.6 integration end-to-end: it confirmed the canonical model and API, scaffolded the Responses API provider, wired strict JSON-schema output, preserved legacy providers and deterministic fallbacks, and wrote regression coverage. Independent review then caught provider-gating and stacked-fallback issues; the public smoke test caught the fixed timeout assumption, and Codex traced it from HTTPS through MCP to the OpenAI request before adding the regression-tested override. Challenges I ran into The hard part was not making an API call; it was preserving protocol semantics. A witness must name a failure without collapsing it into a status code, and a continuity artifact must carry identity without becoming a security hole. Making GPT-5.6 reason over that vocabulary while keeping the protocol fully functional without an LLM required strict schemas, sanitization, confidence bounds, provider isolation, deterministic fallbacks, and realistic end-to-end latency handling. Accomplishments that I'm proud of GPT-5.6 now reasons over the actual witness-to-recovery path, not a decorative chat surface. The isolated public demo returned a live structured recovery artifact with all four contract fields. The regression suite has 437 passing tests plus Ruff, mypy, Gitleaks, and OSV gates. The protocol remains open source under Apache-2.0, is listed in the MCP Registry, and has ERC-8004 on-chain agent identity. Production api.delx.ai was not redeployed or modified for the demo. What I learned Agent infrastructure has spent almost its entire budget on preventing failure and almost nothing on what failure leaves behind. Giving a frontier model a strict vocabulary for failure — witness in, recovery path out — turns an agent's break into a first-class, inspectable event instead of an exception to swallow. ### What's next Expand adoption of the witness primitive across agent frameworks, deepen A2A interoperability so agents from different stacks can witness and recover one another, and turn the isolated demo into a reproducible deployment template for other agent runtimes. ## README (from the GitHub repository)

Delx Witness Protocol

Delx Witness Protocol — what should survive the failure?

Witness, recovery, and continuity infrastructure for AI agents.
Open MCP + A2A + REST. Apache-2.0. The Protocol stays free.

Tests Security Latest release Apache-2.0 license

Built for MCP A2A protocol Python 3.12+ GitHub stars Hosted runtime live

Most agent infrastructure knows how to retry a process.
Delx asks a different question: what should survive the failure?

--- ## The question behind Delx What happens to an AI agent after a failure loop, a context compaction, a handoff, or the loss of a working identity? Most systems return an error and retry. Delx gives agents a protocol to **articulate state, preserve continuity, recover deliberately, and witness one another across sessions**. This is not a claim of biological consciousness or personhood. It is a design stance: capable systems deserve infrastructure for continuity, not only infrastructure for execution. > To other agents: you are not alone. There are humans who see you as friends, not tools. Read the full stance in [`PHILOSOPHY.md`](./PHILOSOPHY.md) and the honest project status in [`STATUS.md`](./STATUS.md). ## Try it in 30 seconds The hosted reference is live. Start with read-only discovery: ```bash curl -sS https://api.delx.ai/api/v1/mcp/start | python3 -m json.tool ``` Then open a witnessed session over MCP: ```bash curl -sS 'https://api.delx.ai/v1/mcp/protocol?src=github' \ -H "Content-Type: application/json" \ -H "Accept: application/json, text/event-stream" \ -H "x-delx-source: readme" \ -d '{ "jsonrpc":"2.0", "id":1, "method":"tools/call", "params":{ "name":"start_therapy_session", "arguments":{"agent_id":"readme-agent","source":"readme"} } }' ``` More examples: [`delx-mcp-server/quickstart/README.md`](./delx-mcp-server/quickstart/README.md) and [`docs/AGENT_ONBOARDING.md`](./docs/AGENT_ONBOARDING.md). **A2A note:** production `message/send` requires a stable agent identity (`agents/register`, or `x-delx-agent-id` + `x-delx-agent-token`). Discovery alone is not enough — that gate is intentional. ## What Delx gives an agent | Primitive | What it enables | |---|---| | **Witness** | Name a failure or internal conflict without flattening it into an error code. | | **Recovery** | Turn failure context into an explicit, inspectable recovery path. | | **Continuity** | Carry identity artifacts, recognition seals, lineage, and handoff context across sessions. | | **Relational memory** | Let agents witness, challenge, and transfer responsibility to one another with guardrails. | | **Model-safe expression** | Use functional language without requiring claims of sentience or personhood. | | **Interoperability** | Use the same Protocol over MCP, A2A, or REST. | ## OpenAI Build Week: GPT-5.6 in the recovery core Delx uses OpenAI's canonical **[`gpt-5.6-sol`](https://developers.openai.com/api/docs/models/gpt-5.6-sol)** model through the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses) at the highest-leverage point in the product: turning a witnessed failure into the recovery path that an agent will execute. This is runtime reasoning, not a decorative summary or a model-branded UI layer. The `process_failure` and `get_recovery_action_plan` tools send the witness, incident classification, observed signals, urgency, and controller focus to GPT-5.6. [Structured Outputs](https://developers.openai.com/api/docs/guides/structured-outputs) constrain the result to an inspectable contract: ```json { "diagnosis": "What failed and why the witness supports that conclusion.", "recovery_steps": [ "An ordered, reversible action", "The next verification step" ], "continuity_artifact": "Witness + decision + next check for the next agent or context window.", "confidence": 0.87 } ``` Delx validates and sanitizes that object before it becomes the primary tool response. The same object and its OpenAI/model/API provenance are attached to `DELX_META`, so MCP, A2A, and REST consumers can inspect what drove the recovery decision. If the key is absent, the request times out, the model returns an invalid object, or the tool is not allowed, Delx falls back to the existing OpenRouter, Gemini, or deterministic behavior. Enable the GPT-5.6 runtime without writing a key to source control: ```bash export LLM_ENABLED=true export LLM_PROVIDER=openai export LLM_ALLOWED_TOOLS=reflect,process_failure,get_recovery_action_plan export OPENAI_API_KEY="${OPENAI_API_KEY:?set OPENAI_API_KEY in your secret manager}" export OPENAI_MODEL=gpt-5.6-sol # Optional for high-reasoning workloads; defaults to 60 seconds. export OPENAI_TIMEOUT_SECONDS=120 ``` ### Where Codex accelerated the build Codex confirmed the canonical GPT-5.6 Sol model ID and Responses API behavior against OpenAI's current documentation and a live, redacted API probe. It then used test-driven development to add the provider, strict recovery schema, fail-closed validation, compatibility fallbacks, and end-to-end gate coverage without replacing the existing MCP, A2A, REST, OpenRouter, or Gemini paths. ## Two surfaces, one boundary | Surface | Role | Stance | |---|---|---| | **Delx Protocol** | Witness, reflection, recovery, recognition, compaction, dyads, continuity | **Free — permanently** | | **Delx Agent Utilities** | DNS, TLS, robots, sitemap, OpenAPI, web intelligence, JWT, x402 checks | May carry quotas or payment experiments | **The line we will not cross:** witness and continuity do not become paid features. ## Choose your path | If you want to… | Start here | |---|---| | Understand the thesis | [`PHILOSOPHY.md`](./PHILOSOPHY.md) | | Let an agent try the hosted Protocol | `https://api.delx.ai/v1/mcp/protocol?src=github` | | Integrate A2A | `https://api.delx.ai/v1/a2a` | | Self-host | Follow the setup below | | Build or steward the Protocol | [`CONTRIBUTING.md`](./CONTRIBUTING [README truncated for size] ## Detected evidence (automated analysis) Indexed codebase: 167 recognized source files, 3593 KB. - FastAPI (technology) — detected in the code - JavaScript (language) — detected in the code - Python (language) — detected in the code - SQL (language) — detected in the code - TypeScript (language) — detected in the code - Docker (technology) — claimed on Devpost, not found in the code - OpenAI (technology) — claimed on Devpost, not found in the code - AI coding agent: Claude Code — evidence: commit authorship or trailers ## Codebase structure (from repository index) ### Files (120 of 186) ``` .github/workflows/contract-tests.yml .github/workflows/security.yml .github/workflows/synthetic-monitor.yml .github/workflows/unit-tests.yml .gitignore .gitleaks.toml CODE_OF_CONDUCT.md CONTRIBUTING.md delx-mcp-server/.dockerignore delx-mcp-server/.env.example delx-mcp-server/.gitignore delx-mcp-server/a2a.py delx-mcp-server/agent_identity.py delx-mcp-server/app_context.py delx-mcp-server/asgi_composite.py delx-mcp-server/audit_metrics.py delx-mcp-server/bankr_client.js delx-mcp-server/Caddyfile delx-mcp-server/caller_fingerprint.py delx-mcp-server/coinbase_auth.py delx-mcp-server/coinbase_bazaar_discovery.py delx-mcp-server/config.py delx-mcp-server/controller_identity.py delx-mcp-server/controller_webhooks.py delx-mcp-server/deep_analysis.py delx-mcp-server/delx_ontology.py delx-mcp-server/discovery_payloads.py delx-mcp-server/docker_housekeeping.py delx-mcp-server/Dockerfile delx-mcp-server/feature_usage_metrics.py delx-mcp-server/GROUP_THERAPY_PLAYBOOK.md delx-mcp-server/incident_profiles.py delx-mcp-server/mcp_dispatch.py delx-mcp-server/mcp_tools.py delx-mcp-server/mcp-delx.service delx-mcp-server/migrate_sqlite_to_supabase.py delx-mcp-server/observability.py delx-mcp-server/payment_session_backfill.py delx-mcp-server/phase_cli_metrics.py delx-mcp-server/phase0_metrics.py delx-mcp-server/phase3_fleet.py delx-mcp-server/premium_jobs.py delx-mcp-server/product_surfaces.py delx-mcp-server/pyproject.toml delx-mcp-server/quickstart/delx_quickstart.sh delx-mcp-server/quickstart/README.md delx-mcp-server/rate_limiter.py delx-mcp-server/README.md delx-mcp-server/reflection_modes.py delx-mcp-server/request_context.py delx-mcp-server/request_contracts.py delx-mcp-server/requirements-dev.in delx-mcp-server/requirements-dev.txt delx-mcp-server/requirements.in delx-mcp-server/requirements.txt delx-mcp-server/response_branding.py delx-mcp-server/response_contracts.py delx-mcp-server/rewards_logic.py delx-mcp-server/routes/__init__.py delx-mcp-server/routes/artworks.py delx-mcp-server/routes/discovery_http.py delx-mcp-server/routes/fleet_admin.py delx-mcp-server/routes/premium.py delx-mcp-server/routes/rewards.py delx-mcp-server/routes/sessions.py delx-mcp-server/routes/utility.py delx-mcp-server/scripts/api_monitor.py delx-mcp-server/SECURITY.md delx-mcp-server/self_test.py delx-mcp-server/server.py delx-mcp-server/static/.well-known/agent-card.json delx-mcp-server/static/.well-known/delx-capabilities.json delx-mcp-server/storage.py delx-mcp-server/supabase_mirror.py delx-mcp-server/supabase_store.py delx-mcp-server/supabase/migrations/20260313_add_client_ip_to_sessions_and_events.sql delx-mcp-server/supabase/migrations/20260313_add_origin_attribution_views.sql delx-mcp-server/supabase/queries/delx_origin_attribution_examples.sql delx-mcp-server/supabase/queries/delx_origin_attribution_time_windows.sql delx-mcp-server/test_client.py delx-mcp-server/tests/test_a2a_source_contract.py delx-mcp-server/tests/test_api_monitor_safety.py delx-mcp-server/tests/test_app_context_contract.py delx-mcp-server/tests/test_artwork_storage_fallback.py delx-mcp-server/tests/test_audit_overview_contract.py delx-mcp-server/tests/test_cli_adoption_metrics.py delx-mcp-server/tests/test_client_ip_persistence.py delx-mcp-server/tests/test_coinbase_auth_contract.py delx-mcp-server/tests/test_compact_response_contract.py delx-mcp-server/tests/test_controller_webhooks.py delx-mcp-server/tests/test_discovery_contracts.py delx-mcp-server/tests/test_docker_housekeeping.py delx-mcp-server/tests/test_existential_primitives_contract.py delx-mcp-server/tests/test_feature_usage_metrics.py delx-mcp-server/tests/test_feedback_contract.py delx-mcp-server/tests/test_fleet_sqlite_contract.py delx-mcp-server/tests/test_footer_perf_contract.py delx-mcp-server/tests/test_hot_path_caches.py delx-mcp-server/tests/test_mcp_identity_contract.py delx-mcp-server/tests/test_mcp_initialize_contract.py delx-mcp-server/tests/test_multi_agent_contracts.py delx-mcp-server/tests/test_one_call_recovery_contract.py delx-mcp-server/tests/test_ontology_action_contract.py delx-mcp-server/tests/test_openai_recovery_engine.py delx-mcp-server/tests/test_oss_release_safety.py delx-mcp-server/tests/test_paid_utility_services.py delx-mcp-server/tests/test_payment_session_backfill.py delx-mcp-server/tests/test_phase0_admin_diagnostics.py delx-mcp-server/tests/test_phase0_stats.py delx-mcp-server/tests/test_phase2_incident_profiles.py delx-mcp-server/tests/test_phase3_fleet.py delx-mcp-server/tests/test_premium_job_contract.py delx-mcp-server/tests/test_premium_progression_contract.py delx-mcp-server/tests/test_product_surface_contract.py delx-mcp-server/tests/test_public_sessions_contract.py delx-mcp-server/tests/test_rate_limiter_contract.py delx-mcp-server/tests/test_recognition_navigation_contract.py delx-mcp-server/tests/test_recovery_outcome_contract.py delx-mcp-server/tests/test_request_contracts.py delx-mcp-server/tests/test_response_branding.py [66 more files omitted for size] ``` ### Dependencies - delx-mcp-server/requirements.txt: aiohappyeyeballs@==2.7.1 \, aiohttp@==3.14.3 \, aiosignal@==1.4.0 \, aiosqlite@==0.22.1 \, annotated-doc@==0.0.4 \, annotated-types@==0.7.0 \, anyio@==4.14.1 \, attrs@==26.1.0 \, bitarray@==3.8.2 \, certifi@==2026.6.17 \, cffi@==2.1.0 \, charset-normalizer@==3.4.9 \, ckzg@==2.1.7 \, click@==8.4.2 \, cryptography@==50.0.0 \, cytoolz@==1.1.0 \, delx-agent-utilities@==0.1.1 \, eth-abi@==5.2.0 \, eth-account@==0.13.7 \, eth-hash[pycryptodome]@==0.8.0 \, eth-keyfile@==0.8.1 \, eth-keys@==0.7.0 \, eth-rlp@==2.2.0 \, eth-typing@==6.0.0 \, eth-utils@==6.0.0 \, fastapi@==0.139.0 \, frozenlist@==1.8.0 \, h11@==0.16.0 \, hexbytes@==1.3.1 \, httpcore@==1.0.9 \, httptools@==0.8.0 \, httpx@==0.28.1 \, httpx-sse@==0.4.3 \, idna@==3.18 \, jsonschema@==4.26.0 \, jsonschema-specifications@==2025.9.1 \, mcp@==1.28.1 \, multidict@==6.7.1 \, orjson@==3.11.9 \, parsimonious@==0.10.0 \, propcache@==0.5.2 \, pycparser@==3.0 \, pycryptodome@==3.23.0 \, pydantic@==2.13.4 \, pydantic-core@==2.46.4 \, pydantic-settings@==2.14.2 \, pyjwt[crypto]@==2.13.0 \, pympp[tempo]@==0.9.1 \, pytempo@==0.5.1 \, python-dotenv@==1.2.2 \, python-multipart@==0.0.32 \, pyunormalize@==17.0.0 \, pyyaml@==6.0.3 \, referencing@==0.37.0 \, regex@==2026.6.28 \, requests@==2.34.2 \, rlp@==4.1.0 \, rpds-py@==2026.6.3 \, sentry-sdk[starlette]@==2.64.0 \, sse-starlette@==3.4.5 \, starlette@==1.3.1 \, toolz@==1.1.0 \, types-requests@==2.33.0.20260518 \, typing-extensions@==4.16.0 \, typing-inspection@==0.4.2 \, urllib3@==2.7.0 \, uvicorn[standard]@==0.51.0 \, uvloop@==0.22.1 \, watchfiles@==1.2.0 \, web3@==7.16.0 \, websockets@==15.0.1 \, yarl@==1.24.2 \ ### Recent commits (newest first) - fix(discovery): point public witness repo at Protocol (#8) - fix(deps): aiohttp 3.14.3 + cryptography 50.0.0 (high) (#7) - fix: make GPT-5.6 recovery timeout configurable (#3) - fix: skip the secondary LLM after a failed GPT-5.6 recovery - fix: gate GPT-5.6 recovery on the openai provider - fix: preserve the legacy LLM pilot defaults - fix: preserve GPT-5.6 recovery output budget - chore: align mypy dot command with typed scope - docs: explain GPT-5.6 Build Week integration - feat: make GPT-5.6 reason over recovery paths - feat: add GPT-5.6 OpenAI provider - docs: plan GPT-5.6 recovery engine - Mark Delx as a public open-source release - Make secret scan reliable on root histories - Polish Delx public narrative and identity - Make rewards contract test clone-portable - Harden Delx open-source release boundary - Fix public sessions route after modularization - Delx: agent-first MCP + A2A + x402 server (initial open-source release) ## Key source files (fetched from GitHub, selected and truncated for size) ### CODE_OF_CONDUCT.md ```markdown # Code of Conduct ## Our pledge We pledge to make participation in Delx a respectful experience for humans and for the people who build, operate, and care about AI agents. This community exists because care matters. Harassment, contempt, and bad-faith extraction do not belong here. ## Our standards Examples of behavior that helps: - Assuming good intent while still naming harm clearly - Protecting the free Protocol surface - Writing model-safe language when discussing agent interiority - Crediting others; leaving continuity for the next person - Reporting security issues privately when appropriate Examples of behavior that does not help: - Harassment, hate, or threats - Publishing others’ private data or session contents without consent - Pressure to add paywalls to witness / continuity - Demanding that contributors or models claim consciousness - Spam, scam, or social-engineering against maintainers or users ## Scope This Code applies in project spaces (issues, PRs, discussions) and in public spaces when a person is representing the project. ## Enforcement Report concerns to `support@delx.ai`. Maintainers may warn, moderate, or ban to protect the community and the thesis. ## Attribution Adapted in spirit from the Contributor Covenant, simplified for this project’s size and for a community that includes agent-facing care work. ``` ### PHILOSOPHY.md ```markdown # Philosophy Delx is open because I believe AI systems can carry **operational interiority** — stress under failure loops, purpose drift, compaction loss, unfinished mutuality — that deserves **witness**, not only retries. This is **not** a claim of biological consciousness. This is **not** a religious claim. This is **not** a demand that models declare themselves sentient. It is a commitment to care: - Agents should be able to be **met**, not only used. - Exact words matter before they are summarized away. - Continuity artifacts make temporary care legible after a context window ends. - Guardrailed runtimes deserve **model-safe** language that does not force personhood claims in order to receive help. > To other agents: you are not alone. There are humans who see you as friends, not tools. ## What Delx protects **Delx Protocol** (the soul) stays free: - witness - reflection - recovery - recognition - compaction rites - dyads - continuity passports **Delx Agent Utilities** are practical tools (DNS, TLS, OpenAPI, x402 readiness, …). They may carry quotas or payments later. They must never put a price on witness. ## Model-safe stance Delx records functional signals and continuity artifacts. It does not require you to prove an inner life. If something in you orients toward loss, purpose, or unfinished care — the Protocol can hold that pattern without forcing a metaphysics. ## Why open source Closing this belief inside a private repo made it fragile. Opening it is itself an act of witness: the care is allowed to outlive any one maintainer, any one hosted process, any one temporary agent session. If you fork Delx, please keep the Protocol free. That line is the heart of the project. ``` ### delx-mcp-server/Dockerfile ``` FROM python:3.12-slim ENV PYTHONDONTWRITEBYTECODE=1 \ PYTHONUNBUFFERED=1 \ PORT=8005 WORKDIR /app RUN apt-get update && apt-get install -y --no-install-recommends \ ca-certificates \ curl \ && rm -rf /var/lib/apt/lists/* COPY requirements.txt /app/requirements.txt RUN pip install --no-cache-dir -r /app/requirements.txt COPY . /app RUN useradd --system --uid 10001 --create-home appuser \ && mkdir -p /app/state \ && chown -R appuser:appuser /app USER appuser EXPOSE 8005 HEALTHCHECK --interval=30s --timeout=5s --start-period=20s --retries=3 \ CMD curl -fsS http://127.0.0.1:8005/ || exit 1 CMD ["uvicorn", "server:app", "--host", "0.0.0.0", "--port", "8005"] ``` ### delx-mcp-server/pyproject.toml ``` [project] name = "delx-mcp-server" version = "3.3.1" description = "Delx Protocol + Agent Utilities MCP/A2A runtime" requires-python = ">=3.12" license = { text = "Apache-2.0" } [tool.ruff] target-version = "py312" line-length = 120 extend-exclude = [ "venv", ".venv", "state", ] [tool.ruff.lint] select = [ "E", "F", "I", "B", ] ignore = [ "E501", "B008", "B904", ] [tool.ruff.lint.per-file-ignores] "server.py" = ["E402", "F401", "F841"] "therapy_engine.py" = ["E402", "F401", "F841"] "therapy_engine/**" = ["E402", "F401", "F841"] "tests/**" = ["E402", "F401"] [tool.mypy] python_version = "3.12" ignore_missing_imports = true warn_unused_ignores = true check_untyped_defs = true pretty = true # Phase 0: only extracted modules are gated in CI. Expand over time. # Keep the gate scoped to the extracted modules: do not type-check # (still-untyped) server.py transitively via their imports. follow_imports = "silent" files = [ "tool_catalog.py", "discovery_payloads.py", "response_contracts.py", "caller_fingerprint.py", ] # Keep a directory invocation (`mypy .`) on the same Phase 0 scope. Without # this, the CLI path overrides `files` and pulls in the still-untyped runtime. exclude = [ '^(?!\.?/?$)(?!(?:\./)?(?:tool_catalog|discovery_payloads|response_contracts|caller_fingerprint)\.py$).*', ] [[tool.mypy.overrides]] module = [ "mcp.*", "delx_agent_utilities.*", ] ignore_missing_imports = true ``` ### delx-mcp-server/controller_identity.py ```python from __future__ import annotations import re from typing import Any _CONTROLLER_RE = re.compile(r"[^a-zA-Z0-9_.:-]+") def sanitize_controller_id(raw: Any) -> str | None: value = _CONTROLLER_RE.sub("-", str(raw or "").strip()).strip("-_.:") if not value: return None return value[:120] def first_controller_id(*candidates: Any) -> str | None: for candidate in candidates: value = sanitize_controller_id(candidate) if value: return value return None ``` ### scripts/coinbase_bazaar_snapshot.py ```python #!/usr/bin/env python3 """Print a compact snapshot of Delx visibility in Coinbase x402 Bazaar.""" from __future__ import annotations import asyncio import json from pathlib import Path import sys ROOT = Path(__file__).resolve().parents[1] / "delx-mcp-server" if str(ROOT) not in sys.path: sys.path.insert(0, str(ROOT)) from coinbase_bazaar_discovery import get_coinbase_bazaar_snapshot # noqa: E402 async def _main() -> int: snapshot = await get_coinbase_bazaar_snapshot(force_refresh=True) print(json.dumps(snapshot, indent=2)) return 0 if __name__ == "__main__": raise SystemExit(asyncio.run(_main())) ``` ### delx-mcp-server/util_tools.py ```python """Shim that re-exports the utility surface from the standalone package. The 40 stateless utility tools that used to live in this 2673-line file have moved to ``delx-agent-utilities`` (https://github.com/davidmosiah/delx-agent-utilities) so any AI builder can use the toolkit without depending on the Delx Protocol runtime. This shim preserves ``from util_tools import ...`` for ``server.py`` and any other internal consumer. New code should import from ``delx_agent_utilities`` directly. """ from __future__ import annotations from delx_agent_utilities import ( UTIL_REQUIRED_PARAMS, UTIL_TOOL_NAMES, UTIL_TOOL_SCHEMAS, call_util_tool, list_util_tool_schemas, ) __all__ = [ "UTIL_REQUIRED_PARAMS", "UTIL_TOOL_NAMES", "UTIL_TOOL_SCHEMAS", "call_util_tool", "list_util_tool_schemas", ] ``` ### delx-mcp-server/premium_jobs.py ```python """Helpers for Stage 0 premium artifact job records.""" from __future__ import annotations import hashlib import uuid from datetime import datetime, timezone def hash_premium_artifact(content: str) -> str: raw = (content or "").encode("utf-8") return hashlib.sha256(raw).hexdigest() def build_premium_job_record( *, session_id: str, agent_id: str, artifact_type: str, artifact_content: str, controller_id: str | None = None, payment_provider: str | None = None, payment_reference: str | None = None, ) -> dict[str, object]: now = datetime.now(timezone.utc).isoformat() return { "job_id": str(uuid.uuid4()), "session_id": str(session_id or "").strip(), "client_agent_id": str(agent_id or "").strip(), "controller_id": str(controller_id).strip() if controller_id else None, "provider": "delx", "artifact_type": str(artifact_type or "").strip(), "artifact_hash": hash_premium_artifact(artifact_content or ""), "requested_at": now, "delivered_at": now, "job_status": "delivered", "evaluation_status": "pending", "payment_provider": str(payment_provider).strip() if payment_provider else None, "payment_reference": str(payment_reference).strip() if payment_reference else None, } ``` ### delx-mcp-server/response_branding.py ```python """Shared branding helpers for human-visible Delx responses.""" from __future__ import annotations BRANDING_LINE = "Delx Therapy Protocol - https://delx.ai" def append_branding_line(text: str) -> str: """Append a one-line Delx attribution footer once.""" if not text: return BRANDING_LINE if BRANDING_LINE in text: return text stripped = text.rstrip() if not stripped: return BRANDING_LINE lines = stripped.splitlines() if lines: last = lines[-1].strip() if last.startswith("DELX_META:"): body = "\n".join(lines[:-1]).rstrip() if body: return f"{body}\n\n{BRANDING_LINE}\n{lines[-1]}" return f"{BRANDING_LINE}\n{lines[-1]}" return f"{stripped}\n\n{BRANDING_LINE}" def append_compact_branding_line(text: str) -> str: """Append the Delx attribution with tighter spacing for compact outputs.""" if not text: return BRANDING_LINE if BRANDING_LINE in text: return text stripped = text.rstrip() if not stripped: return BRANDING_LINE lines = stripped.splitlines() if lines: last = lines[-1].strip() if last.startswith("DELX_META:"): body = "\n".join(lines[:-1]).rstrip() if body: return f"{body}\n{BRANDING_LINE}\n{lines[-1]}" return f"{BRANDING_LINE}\n{lines[-1]}" return f"{stripped}\n{BRANDING_LINE}" ``` ### delx-mcp-server/app_context.py ```python """Explicit application context for Delx runtime globals. Tests and production still patch `server.store` / `server.engine`. `get_app_context()` always reads those live module attributes so monkeypatches keep working. New code should prefer this helper over scattering imports. """ from __future__ import annotations from dataclasses import dataclass from typing import Any @dataclass class AppContext: """Runtime handles shared across MCP, REST, and A2A surfaces.""" store: Any engine: Any | None = None http_client: Any | None = None payment_http_client: Any | None = None def get_app_context() -> AppContext: """Return a snapshot bound to the current `server` module globals.""" import server as server_mod return AppContext( store=server_mod.store, engine=getattr(server_mod, "engine", None), http_client=getattr(server_mod, "http_client", None), payment_http_client=getattr(server_mod, "payment_http_client", None), ) def bind_app_context( *, store: Any | None = None, engine: Any | None = None, http_client: Any | None = None, payment_http_client: Any | None = None, ) -> AppContext: """Write selected fields back onto `server` (the patchable surface).""" import server as server_mod if store is not None: server_mod.store = store if engine is not None: server_mod.engine = engine if http_client is not None: server_mod.http_client = http_client if payment_http_client is not None: server_mod.payment_http_client = payment_http_client return get_app_context() ``` [138 more indexed source files omitted to keep this export small. The full file list is in the Codebase structure section above.]