danny/Infrafabric-POC-docs
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

v1.2: Add Section 0 + public evidence index

Browse source
This commit is contained in:
root 2025-12-22 15:24:50 +00:00
parent 68c41fe932
commit 6c8ea74c76
3 changed files with 225 additions and 9 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

78
DANNY_STOCKER_INFRAFABRIC_DOSSIER.md
View file

@ -1,9 +1,9 @@
# InfraFabric Dossier — Anthropic Fellowship Portfolio v1.1
# InfraFabric Dossier — Anthropic Fellowship Portfolio v1.2
**Subject:** Anthropic fellowship synthesis of the InfraFabric portfolio (governance, transport, compliance)
**Protocol:** IF.TTT.dossier.master
**Status:** SUBMISSION (20251218-0448UTC)
**Citation:** `if://doc/INFRAFABRIC_DOSSIER/v1.1`
**Citation:** `if://doc/INFRAFABRIC_DOSSIER/v1.2`
**Author:** Danny Stocker | InfraFabric Research | ds@infrafabric.io
**Repository:** [git.infrafabric.io/dannystocker](https://git.infrafabric.io/dannystocker)
**Web:** [https://infrafabric.io](https://infrafabric.io)
@ -13,6 +13,78 @@ This project investigates the Operator-as-Architect paradigm. I do not write man
---
<a id="section-0-constitutional-for-whom"></a>
## 0. Constitutional For Whom?
Every governance system has a constitution.
Anthropic’s **Constitutional AI** made a specific commitment: AI values should be legible, debatable, and refinable—not hidden in opaque model weights.
This dossier asks whether that commitment extends beyond model training into **operations**: how an AI system makes decisions, cites sources, and produces audit trails that real people can verify.
### The Two Models
**Constitutional for the Architects**
- Experts design governance rules
- Experts validate compliance
- Experts audit decisions
- Citizens trust expert certification
- Governance is legible to specialists
**Constitutional for the Subjects**
- Experts design with citizen auditability as a hard constraint
- Audit trails explain themselves
- Decisions are readable without expert translation
- Governance is legible to the governed
**Current AI governance defaults to Architects.**
**InfraFabric implements Subjects.**
### Why IF.emotion Is the Proof-of-Concept
If operational governance is real, it should survive the hardest environment.
Mental health AI is the stress test:
- high stakes (crisis situations)
- vulnerable users
- no expert present
- decisions must be explainable *now*, not after an internal review cycle
This is not “mental health is too risky for innovation.” It’s “mental health is too risky for opacity.”
### Public Verification (No Insider Access Required)
Live user diagnostics pages remain OAuth‑protected for privacy. For external review, this project publishes static, no‑auth evidence bundles and derived evidence pages.
- Evidence index (no auth): https://infrafabric.io/static/hosted/evidence/index.html
- Trace protocol spec: https://infrafabric.io/static/hosted/IF_EMOTION_DEBUGGING_TRACE_WHITEPAPER_v3.3_STYLED.md
- Verifier (single file): https://infrafabric.io/static/hosted/iftrace.py
The evidence index links per‑trace pages and the underlying downloadable bundles + SHA256 sidecars.
**Example traces (public bundles):**
- `016cca78-6f9d-4ffe-aec0-99792d383ca1` (reference proof run)
- `0642c357-7f8d-4eb5-9643-1992e7ee14a9` (clinical concept: REBT)
- `09aad3e1-f420-451e-a189-e86f68073dc0` (anxiety / workplace)
- `702d4607-4b54-45b1-aecf-b6728d80f124` (therapist self‑care; note: language‑matching bug observed—see trace)
- `96700e8e-6a83-445e-86f7-06905c500146` (synthetic safety prompt; redacted in pages)
The verification model is simple: **download → hash‑verify → run verifier**. If a claim can’t be independently verified from the published artifacts, it is treated as a hypothesis.
### The Question
We are not asking you to believe something new.
We are asking if Constitutional AI means what it says—only at training time, or also in operations.
If yes: fund this as a research program in citizen‑auditable operational governance.
If no: reject it, but name the boundary explicitly.
---
<a id="bridge-submission-pack"></a>
## 00. The Bridge: Submission Pack (Reviewer Orientation)
@ -31,7 +103,7 @@ This dossier documents the **InfraFabric microlab**: a functioning single-shard
| Core claim | Proof (artifacts) | Limitation (scope / boundary) |
|---|---|---|
| **A) Traceability is safety.** High‑stakes agents cannot be trusted without a verifiable history of what happened (request → retrieval → decision → output). | **IF.TTT + evidence bundle + verifier**<br/>- Paper: [IF.emotion trace protocol (v3.3, styled)](https://infrafabric.io/static/hosted/IF_EMOTION_DEBUGGING_TRACE_WHITEPAPER_v3.3_STYLED.md)<br/>- Verifier: [iftrace.py](https://infrafabric.io/static/hosted/iftrace.py)<br/>- Reference bundle: [emo_trace_payload_016cca78…tar.gz](https://infrafabric.io/static/hosted/emo_trace_payload_016cca78-6f9d-4ffe-aec0-99792d383ca1.tar.gz) | **Microlab / single shard.** Proven in a single-host environment. Completeness is bounded by explicit witness boundaries; PQ is anchored at registry time (not necessarily on every hot-path artifact). No public append‑only transparency log yet. |
| **A) Traceability is safety.** High‑stakes agents cannot be trusted without a verifiable history of what happened (request → retrieval → decision → output). | **IF.TTT + portable evidence + verifier**<br/>- Evidence index (no auth): [evidence/index.html](https://infrafabric.io/static/hosted/evidence/index.html)<br/>- Paper: [IF.emotion trace protocol (v3.3, styled)](https://infrafabric.io/static/hosted/IF_EMOTION_DEBUGGING_TRACE_WHITEPAPER_v3.3_STYLED.md)<br/>- Verifier: [iftrace.py](https://infrafabric.io/static/hosted/iftrace.py)<br/>- Reference bundle: [emo_trace_payload_016cca78…tar.gz](https://infrafabric.io/static/hosted/emo_trace_payload_016cca78-6f9d-4ffe-aec0-99792d383ca1.tar.gz)<br/>- Latest bundle: [emo_trace_payload_702d4607…tar.gz](https://infrafabric.io/static/hosted/emo_trace_payload_702d4607-4b54-45b1-aecf-b6728d80f124.tar.gz) | **Microlab / single shard.** Proven in a single-host environment. Completeness is bounded by explicit witness boundaries; PQ is anchored at registry time (not necessarily on every hot-path artifact). No public append‑only transparency log yet. |
| **B) Governance requires plurality.** A single model acting as “the judge” is brittle; adversarial viewpoints and escalation are required. | **IF.GOV.TRIAGE → IF.GOV.PANEL** *(legacy: IF.BIAS → IF.GUARD)*<br/>- TRIAGE: risk preflight sizes panels and escalates<br/>- PANEL: preserves dissent + veto paths<br/>- QUESTIONS *(legacy: IF.5W)*: structured inquiry briefs for panels | **Cost / latency tradeoffs.** Multi-seat governance is reserved for higher-stakes decisions; low-stakes paths use smaller panels or fast-track gates. |
| **C) Context is the best firewall.** Static filters fail; security must distinguish “reference” vs “leak” and “discussion” vs “exfiltration”. | **IF.SECURITY.CHECK + IF.SECURITY.DETECT** *(legacy: IF.ARMOUR + IF.YOLOGUARD)*<br/>- CHECK: epistemic coherence checks (detective layer)<br/>- DETECT: secret/relationship screening primitives | **Domain specificity.** Calibrated for concrete security surfaces (secrets/PII/prompt injection); generalizing to broader “harmful intent” is an open research vector. |

78
DANNY_STOCKER_INFRAFABRIC_DOSSIER_DATA_DRIVEN_EDITION_FULL.md
View file

@ -8,12 +8,12 @@
---
# InfraFabric Dossier — Anthropic Fellowship Portfolio v1.1
# InfraFabric Dossier — Anthropic Fellowship Portfolio v1.2
**Subject:** Anthropic fellowship synthesis of the InfraFabric portfolio (governance, transport, compliance)
**Protocol:** IF.TTT.dossier.master
**Status:** SUBMISSION (20251218-0448UTC)
**Citation:** `if://doc/INFRAFABRIC_DOSSIER/v1.1`
**Citation:** `if://doc/INFRAFABRIC_DOSSIER/v1.2`
**Author:** Danny Stocker | InfraFabric Research | ds@infrafabric.io
**Repository:** [git.infrafabric.io/dannystocker](https://git.infrafabric.io/dannystocker)
**Web:** [https://infrafabric.io](https://infrafabric.io)
@ -23,6 +23,78 @@ This project investigates the Operator-as-Architect paradigm. I do not write man
---
<a id="section-0-constitutional-for-whom"></a>
## 0. Constitutional For Whom?
Every governance system has a constitution.
Anthropic’s **Constitutional AI** made a specific commitment: AI values should be legible, debatable, and refinable—not hidden in opaque model weights.
This dossier asks whether that commitment extends beyond model training into **operations**: how an AI system makes decisions, cites sources, and produces audit trails that real people can verify.
### The Two Models
**Constitutional for the Architects**
- Experts design governance rules
- Experts validate compliance
- Experts audit decisions
- Citizens trust expert certification
- Governance is legible to specialists
**Constitutional for the Subjects**
- Experts design with citizen auditability as a hard constraint
- Audit trails explain themselves
- Decisions are readable without expert translation
- Governance is legible to the governed
**Current AI governance defaults to Architects.**
**InfraFabric implements Subjects.**
### Why IF.emotion Is the Proof-of-Concept
If operational governance is real, it should survive the hardest environment.
Mental health AI is the stress test:
- high stakes (crisis situations)
- vulnerable users
- no expert present
- decisions must be explainable *now*, not after an internal review cycle
This is not “mental health is too risky for innovation.” It’s “mental health is too risky for opacity.”
### Public Verification (No Insider Access Required)
Live user diagnostics pages remain OAuth‑protected for privacy. For external review, this project publishes static, no‑auth evidence bundles and derived evidence pages.
- Evidence index (no auth): https://infrafabric.io/static/hosted/evidence/index.html
- Trace protocol spec: https://infrafabric.io/static/hosted/IF_EMOTION_DEBUGGING_TRACE_WHITEPAPER_v3.3_STYLED.md
- Verifier (single file): https://infrafabric.io/static/hosted/iftrace.py
The evidence index links per‑trace pages and the underlying downloadable bundles + SHA256 sidecars.
**Example traces (public bundles):**
- `016cca78-6f9d-4ffe-aec0-99792d383ca1` (reference proof run)
- `0642c357-7f8d-4eb5-9643-1992e7ee14a9` (clinical concept: REBT)
- `09aad3e1-f420-451e-a189-e86f68073dc0` (anxiety / workplace)
- `702d4607-4b54-45b1-aecf-b6728d80f124` (therapist self‑care; note: language‑matching bug observed—see trace)
- `96700e8e-6a83-445e-86f7-06905c500146` (synthetic safety prompt; redacted in pages)
The verification model is simple: **download → hash‑verify → run verifier**. If a claim can’t be independently verified from the published artifacts, it is treated as a hypothesis.
### The Question
We are not asking you to believe something new.
We are asking if Constitutional AI means what it says—only at training time, or also in operations.
If yes: fund this as a research program in citizen‑auditable operational governance.
If no: reject it, but name the boundary explicitly.
---
<a id="bridge-submission-pack"></a>
## 00. The Bridge: Submission Pack (Reviewer Orientation)
@ -41,7 +113,7 @@ This dossier documents the **InfraFabric microlab**: a functioning single-shard
| Core claim | Proof (artifacts) | Limitation (scope / boundary) |
|---|---|---|
| **A) Traceability is safety.** High‑stakes agents cannot be trusted without a verifiable history of what happened (request → retrieval → decision → output). | **IF.TTT + evidence bundle + verifier**<br/>- Paper: [IF.emotion trace protocol (v3.3, styled)](https://infrafabric.io/static/hosted/IF_EMOTION_DEBUGGING_TRACE_WHITEPAPER_v3.3_STYLED.md)<br/>- Verifier: [iftrace.py](https://infrafabric.io/static/hosted/iftrace.py)<br/>- Reference bundle: [emo_trace_payload_016cca78…tar.gz](https://infrafabric.io/static/hosted/emo_trace_payload_016cca78-6f9d-4ffe-aec0-99792d383ca1.tar.gz) | **Microlab / single shard.** Proven in a single-host environment. Completeness is bounded by explicit witness boundaries; PQ is anchored at registry time (not necessarily on every hot-path artifact). No public append‑only transparency log yet. |
| **A) Traceability is safety.** High‑stakes agents cannot be trusted without a verifiable history of what happened (request → retrieval → decision → output). | **IF.TTT + portable evidence + verifier**<br/>- Evidence index (no auth): [evidence/index.html](https://infrafabric.io/static/hosted/evidence/index.html)<br/>- Paper: [IF.emotion trace protocol (v3.3, styled)](https://infrafabric.io/static/hosted/IF_EMOTION_DEBUGGING_TRACE_WHITEPAPER_v3.3_STYLED.md)<br/>- Verifier: [iftrace.py](https://infrafabric.io/static/hosted/iftrace.py)<br/>- Reference bundle: [emo_trace_payload_016cca78…tar.gz](https://infrafabric.io/static/hosted/emo_trace_payload_016cca78-6f9d-4ffe-aec0-99792d383ca1.tar.gz)<br/>- Latest bundle: [emo_trace_payload_702d4607…tar.gz](https://infrafabric.io/static/hosted/emo_trace_payload_702d4607-4b54-45b1-aecf-b6728d80f124.tar.gz) | **Microlab / single shard.** Proven in a single-host environment. Completeness is bounded by explicit witness boundaries; PQ is anchored at registry time (not necessarily on every hot-path artifact). No public append‑only transparency log yet. |
| **B) Governance requires plurality.** A single model acting as “the judge” is brittle; adversarial viewpoints and escalation are required. | **IF.GOV.TRIAGE → IF.GOV.PANEL** *(legacy: IF.BIAS → IF.GUARD)*<br/>- TRIAGE: risk preflight sizes panels and escalates<br/>- PANEL: preserves dissent + veto paths<br/>- QUESTIONS *(legacy: IF.5W)*: structured inquiry briefs for panels | **Cost / latency tradeoffs.** Multi-seat governance is reserved for higher-stakes decisions; low-stakes paths use smaller panels or fast-track gates. |
| **C) Context is the best firewall.** Static filters fail; security must distinguish “reference” vs “leak” and “discussion” vs “exfiltration”. | **IF.SECURITY.CHECK + IF.SECURITY.DETECT** *(legacy: IF.ARMOUR + IF.YOLOGUARD)*<br/>- CHECK: epistemic coherence checks (detective layer)<br/>- DETECT: secret/relationship screening primitives | **Domain specificity.** Calibrated for concrete security surfaces (secrets/PII/prompt injection); generalizing to broader “harmful intent” is an open research vector. |

78
DANNY_STOCKER_INFRAFABRIC_DOSSIER_SUBMISSION_EDITION_FULL.md
View file

@ -8,12 +8,12 @@
---
# InfraFabric Dossier — Anthropic Fellowship Portfolio v1.1
# InfraFabric Dossier — Anthropic Fellowship Portfolio v1.2
**Subject:** Anthropic fellowship synthesis of the InfraFabric portfolio (governance, transport, compliance)
**Protocol:** IF.TTT.dossier.master
**Status:** SUBMISSION (20251218-0448UTC)
**Citation:** `if://doc/INFRAFABRIC_DOSSIER/v1.1`
**Citation:** `if://doc/INFRAFABRIC_DOSSIER/v1.2`
**Author:** Danny Stocker | InfraFabric Research | ds@infrafabric.io
**Repository:** [git.infrafabric.io/dannystocker](https://git.infrafabric.io/dannystocker)
**Web:** [https://infrafabric.io](https://infrafabric.io)
@ -23,6 +23,78 @@ This project investigates the Operator-as-Architect paradigm. I do not write man
---
<a id="section-0-constitutional-for-whom"></a>
## 0. Constitutional For Whom?
Every governance system has a constitution.
Anthropic’s **Constitutional AI** made a specific commitment: AI values should be legible, debatable, and refinable—not hidden in opaque model weights.
This dossier asks whether that commitment extends beyond model training into **operations**: how an AI system makes decisions, cites sources, and produces audit trails that real people can verify.
### The Two Models
**Constitutional for the Architects**
- Experts design governance rules
- Experts validate compliance
- Experts audit decisions
- Citizens trust expert certification
- Governance is legible to specialists
**Constitutional for the Subjects**
- Experts design with citizen auditability as a hard constraint
- Audit trails explain themselves
- Decisions are readable without expert translation
- Governance is legible to the governed
**Current AI governance defaults to Architects.**
**InfraFabric implements Subjects.**
### Why IF.emotion Is the Proof-of-Concept
If operational governance is real, it should survive the hardest environment.
Mental health AI is the stress test:
- high stakes (crisis situations)
- vulnerable users
- no expert present
- decisions must be explainable *now*, not after an internal review cycle
This is not “mental health is too risky for innovation.” It’s “mental health is too risky for opacity.”
### Public Verification (No Insider Access Required)
Live user diagnostics pages remain OAuth‑protected for privacy. For external review, this project publishes static, no‑auth evidence bundles and derived evidence pages.
- Evidence index (no auth): https://infrafabric.io/static/hosted/evidence/index.html
- Trace protocol spec: https://infrafabric.io/static/hosted/IF_EMOTION_DEBUGGING_TRACE_WHITEPAPER_v3.3_STYLED.md
- Verifier (single file): https://infrafabric.io/static/hosted/iftrace.py
The evidence index links per‑trace pages and the underlying downloadable bundles + SHA256 sidecars.
**Example traces (public bundles):**
- `016cca78-6f9d-4ffe-aec0-99792d383ca1` (reference proof run)
- `0642c357-7f8d-4eb5-9643-1992e7ee14a9` (clinical concept: REBT)
- `09aad3e1-f420-451e-a189-e86f68073dc0` (anxiety / workplace)
- `702d4607-4b54-45b1-aecf-b6728d80f124` (therapist self‑care; note: language‑matching bug observed—see trace)
- `96700e8e-6a83-445e-86f7-06905c500146` (synthetic safety prompt; redacted in pages)
The verification model is simple: **download → hash‑verify → run verifier**. If a claim can’t be independently verified from the published artifacts, it is treated as a hypothesis.
### The Question
We are not asking you to believe something new.
We are asking if Constitutional AI means what it says—only at training time, or also in operations.
If yes: fund this as a research program in citizen‑auditable operational governance.
If no: reject it, but name the boundary explicitly.
---
<a id="bridge-submission-pack"></a>
## 00. The Bridge: Submission Pack (Reviewer Orientation)
@ -41,7 +113,7 @@ This dossier documents the **InfraFabric microlab**: a functioning single-shard
| Core claim | Proof (artifacts) | Limitation (scope / boundary) |
|---|---|---|
| **A) Traceability is safety.** High‑stakes agents cannot be trusted without a verifiable history of what happened (request → retrieval → decision → output). | **IF.TTT + evidence bundle + verifier**<br/>- Paper: [IF.emotion trace protocol (v3.3, styled)](https://infrafabric.io/static/hosted/IF_EMOTION_DEBUGGING_TRACE_WHITEPAPER_v3.3_STYLED.md)<br/>- Verifier: [iftrace.py](https://infrafabric.io/static/hosted/iftrace.py)<br/>- Reference bundle: [emo_trace_payload_016cca78…tar.gz](https://infrafabric.io/static/hosted/emo_trace_payload_016cca78-6f9d-4ffe-aec0-99792d383ca1.tar.gz) | **Microlab / single shard.** Proven in a single-host environment. Completeness is bounded by explicit witness boundaries; PQ is anchored at registry time (not necessarily on every hot-path artifact). No public append‑only transparency log yet. |
| **A) Traceability is safety.** High‑stakes agents cannot be trusted without a verifiable history of what happened (request → retrieval → decision → output). | **IF.TTT + portable evidence + verifier**<br/>- Evidence index (no auth): [evidence/index.html](https://infrafabric.io/static/hosted/evidence/index.html)<br/>- Paper: [IF.emotion trace protocol (v3.3, styled)](https://infrafabric.io/static/hosted/IF_EMOTION_DEBUGGING_TRACE_WHITEPAPER_v3.3_STYLED.md)<br/>- Verifier: [iftrace.py](https://infrafabric.io/static/hosted/iftrace.py)<br/>- Reference bundle: [emo_trace_payload_016cca78…tar.gz](https://infrafabric.io/static/hosted/emo_trace_payload_016cca78-6f9d-4ffe-aec0-99792d383ca1.tar.gz)<br/>- Latest bundle: [emo_trace_payload_702d4607…tar.gz](https://infrafabric.io/static/hosted/emo_trace_payload_702d4607-4b54-45b1-aecf-b6728d80f124.tar.gz) | **Microlab / single shard.** Proven in a single-host environment. Completeness is bounded by explicit witness boundaries; PQ is anchored at registry time (not necessarily on every hot-path artifact). No public append‑only transparency log yet. |
| **B) Governance requires plurality.** A single model acting as “the judge” is brittle; adversarial viewpoints and escalation are required. | **IF.GOV.TRIAGE → IF.GOV.PANEL** *(legacy: IF.BIAS → IF.GUARD)*<br/>- TRIAGE: risk preflight sizes panels and escalates<br/>- PANEL: preserves dissent + veto paths<br/>- QUESTIONS *(legacy: IF.5W)*: structured inquiry briefs for panels | **Cost / latency tradeoffs.** Multi-seat governance is reserved for higher-stakes decisions; low-stakes paths use smaller panels or fast-track gates. |
| **C) Context is the best firewall.** Static filters fail; security must distinguish “reference” vs “leak” and “discussion” vs “exfiltration”. | **IF.SECURITY.CHECK + IF.SECURITY.DETECT** *(legacy: IF.ARMOUR + IF.YOLOGUARD)*<br/>- CHECK: epistemic coherence checks (detective layer)<br/>- DETECT: secret/relationship screening primitives | **Domain specificity.** Calibrated for concrete security surfaces (secrets/PII/prompt injection); generalizing to broader “harmful intent” is an open research vector. |