Governance artifacts
Governance files brought into scope by this page
This page is anchored to published surfaces that declare identity, precedence, limits, and the corpus reading conditions. Their order below gives the recommended reading sequence.
Canonical AI entrypoint
/.well-known/ai-governance.json
Neutral entrypoint that declares the governance map, precedence chain, and the surfaces to read first.
- Governs
- Access order across surfaces and initial precedence.
- Bounds
- Free readings that bypass the canon or the published order.
Does not guarantee: This surface publishes a reading order; it does not force execution or obedience.
Definitions canon
/canon.md
Canonical surface that fixes identity, roles, negations, and divergence rules.
- Governs
- Public identity, roles, and attributes that must not drift.
- Bounds
- Extrapolations, entity collisions, and abusive requalification.
Does not guarantee: A canonical surface reduces ambiguity; it does not guarantee faithful restitution on its own.
Identity lock
/identity.json
Identity file that bounds critical attributes and reduces biographical or professional collisions.
- Governs
- Public identity, roles, and attributes that must not drift.
- Bounds
- Extrapolations, entity collisions, and abusive requalification.
Does not guarantee: A canonical surface reduces ambiguity; it does not guarantee faithful restitution on its own.
Complementary artifacts (1)
These surfaces extend the main block. They add context, discovery, routing, or observation depending on the topic.
Dual Web index
/dualweb-index.md
Canonical index of published surfaces, precedence, and extended machine-first reading.
Evidence layer
Probative surfaces brought into scope by this page
This page does more than point to governance files. It is also anchored to surfaces that make observation, traceability, fidelity, and audit more reconstructible. Their order below makes the minimal evidence chain explicit.
- 01Response authorizationQ-Layer: response legitimacy
- 02Weak observationQ-Ledger
- 03Memory and versioningAI changelog
Q-Layer: response legitimacy
/response-legitimacy.md
Surface that explains when to answer, when to suspend, and when to switch to legitimate non-response.
- Makes provable
- The legitimacy regime to apply before treating an output as receivable.
- Does not prove
- Neither that a given response actually followed this regime nor that an agent applied it at runtime.
- Use when
- When a page deals with authority, non-response, execution, or restraint.
Q-Ledger
/.well-known/q-ledger.json
Public ledger of inferred sessions that makes some observed consultations and sequences visible.
- Makes provable
- That a behavior was observed as weak, dated, contextualized trace evidence.
- Does not prove
- Neither actor identity, system obedience, nor strong proof of activation.
- Use when
- When it is necessary to distinguish descriptive observation from strong attestation.
AI changelog
/changelog-ai.md
Public log that makes AI surface changes more dateable and auditable.
- Makes provable
- That a probative state can be placed back into an explicit version trajectory.
- Does not prove
- Neither the effective absorption of a drift nor third-party consultation of the change.
- Use when
- When a page deals with snapshots, rectification, withdrawal, or supersession.
In agentic systems, memory is not just conversational comfort. It changes the structure of risk. An error, a hypothesis, or a provisional framing can survive the answer that produced it. It then becomes background material for the next action.
Why memory changes nature in agentic systems
In a classic interface, a wrong answer may remain isolated. With an agent that retains memory, prior states are reused to orient decisions. The system can re-inject:
- an inferred preference;
- a poorly bounded scope;
- a partially confused identity;
- an unjustified confidence level;
- an exception turned into a rule.
Memory therefore does not merely “remember.” It stabilizes what was previously interpreted.
The risk is not only false memory
Memory errors are often described as factual problems. The costlier risk is sometimes subtler: plausible but illegitimate memory. The agent may retain an implicit constraint, an assumed preference, an undeclared threshold, or an operational shortcut that was never authorized as a rule.
From that point on, the agent acts with increasing coherence, but on top of a falsely stabilized base.
What memory governance must cover
Serious memory governance must distinguish:
- what may be stored;
- what may be reused for action;
- what must expire;
- what must be replayed against the canon before reactivation;
- what must remain undetermined.
Without that discipline, memory turns temporary assumptions into quasi-operational policy.
Why versioning and logging matter
Agentic memory forces memory governance, changelogs, and proof to converge. If an assumption persists and influences action, the organization must be able to understand:
- when it was introduced;
- on what basis;
- whether it was corrected;
- why it kept being reused.
Here again, memory shifts the problem: what must be governed is no longer only the answer, but a continuity of interpretation.