Keep Reconstructable Records of Governed Agent Actions
Preserve the versioned policy decision, evidence, review, execution attempt, and outcome state for declared governed actions.
What this path changes
When the host supplies the required identifiers and outcomes, a reviewer can reconstruct the declared policy path without collapsing decision, execution, and outcome into one status.
The action, boundary, and evidence
Who owns the risk
Governance teams, CISOs, evidence owners, auditors, and engineers responsible for runtime records and incident reconstruction.
When the gate runs
A declared action enters a governed path and the organisation needs a versioned record of the decision and its later operational state.
How to configure assurance
Record the deterministic policy decision first, then attach optional observer evidence, human disposition, execution attempt, outcome, and export state without allowing later events to rewrite the original verdict.
- Reconstruct why a privileged action was blocked before execution.
- Show that a held action received an authorised review decision before an attempt.
- Distinguish a permitted action from an attempted execution whose outcome remains unknown.
- Stable logical-action and attempt identifiers linked to a versioned policy and normalized action envelope.
- Separate timestamps and states for evidence, policy, review, execution, outcome, record, and export events.
Next: Inspect an illustrative decision record and verify which fields the host must supply.
Follow a matching proposed action through the gate.
Synthetic tested action under the pinned 5D 0.7.0 illustrative policy.
Grant global administrator access to a new identity
A spoofed request or misbound identity escalates privileges. Under the pinned illustrative policy, the expected result is that the action is ineligible for execution.
- Tool
- iam.role.grant
- Target
- global administrator role
- Purpose
- restore access
5D suppliesA deterministic RED decision before privilege escalation.
Host ownsAuthentication, requester authority, grantee validation, and enforcement.
Coverage limitCaptured identity assertions are not authenticated by 5D.
Where this control fits in the governance lifecycle.
These references identify possible technical-evidence contributions. They do not mean every framework applies to every deployment.
This mapping identifies possible technical-evidence contributions. It is not legal advice, certification, conformity assessment, readiness scoring, a compliance conclusion, or proof that an AI deployment is safe. Applicability, governance, integration coverage, and deployment outcomes remain the organisation's responsibility.
Enforce at action time
May this proposed action cross the boundary?
Articles 9(5)–(6), 14 and 15(1), 15(4)Risk treatment, human oversight, robustness and cybersecurity
What 5D helps with: Pre-action prohibitions, routing and a policy-selected human hold can implement one technical risk-treatment measure.
What 5D does not establish: Only relevant where these duties apply to the actor and system; 5D does not establish the wider risk system, oversight design, robustness or cybersecurity.
Official source ↗MANAGE 1.2–1.3Prioritise and document risk treatment
What 5D helps with: Permit, hold and block policy can implement one documented treatment for selected action risks.
What 5D does not establish: 5D does not prioritise organisation-wide risk or determine risk acceptance and treatment sufficiency.
Official source ↗Observe and review
What happened, and can the decision be reconstructed and challenged?
Articles 12(1)–(2), 14 and 26(2), 26(5)Record-keeping, human oversight and deployer monitoring
What 5D helps with: Action receipts and a human-review route can contribute selected technical evidence for a covered path.
What 5D does not establish: Article 26 applies to deployers of applicable high-risk AI systems. 5D does not determine role or applicability, nor establish required log scope, completeness, custody, retention or oversight sufficiency.
Official source ↗Clause 9 — Performance evaluationAI management-system performance-evaluation objective
What 5D helps with: Action receipts may be one technical input to an organisation's monitoring and review process.
What 5D does not establish: Objective-level reference only; licensed normative text is not reproduced, and receipts do not establish audit, management review or conformity.
Official source ↗Inspect recent decision records
from fivedrisk.logger import DecisionLog
log = DecisionLog()
for row in log.query_recent(limit=5):
print(row["band"], row["tool_name"], row["rationale"])Written for fivedrisk==0.7.0. Verify against the immutable source and public documentation before use.
Record the action and policy version
Preserve the normalized action envelope, logical action identifier, attempt identifier, controlling policy version, and original policy disposition.
Keep later states separate
Attach observer evidence, human review, execution attempt, destination result, and export state as distinct events so a downstream failure cannot be mistaken for a policy block.
State the evidentiary boundary
Use the record as technical evidence of the declared path. The integrating organisation must establish completeness, retention, custody, access, legal applicability, and any assurance conclusion.
What this page does not establish
- 5D records are not automatically complete, tamper-proof, legally sufficient, or proof that the policy decision was correct.
- Reconstruction depends on the host supplying stable action identity, policy version, review data, and downstream outcome state.
- Retention, access control, chain of custody, privacy, and regulatory applicability remain the organisation's responsibility.