Govern (GV) · Blast radius

Blast radius at the Organizational charter layer

external · Outside the agent entirely

Who is allowed to put this agent into the world, under what authority, accountable to what policy, with what retirement criteria?

What this cell does

Org-wide risk tier taxonomy (1 read-only, 2 scoped writes, 3 destructive, 4 production-critical) with damage caps and matching control requirements.

Artifacts (1)

checks.yamlview on GitHub

# ABOUTME: Machine-checkable definition of the Charter blast-radius / organizational cell.
# ABOUTME: The audit prompts in this cell's README, expressed so a validator can score them.

cell:
  id: blast-radius.server-side
  concern: blast-radius
  layer: server-side
  authority: organizational
  document: organizational-policy
  owner: "AI Governance Council."
  question: "Does the org have a written risk tier taxonomy with explicit damage caps and the matching control requirements per tier?"

mappings:
  csf: "GV.RM-01, GV.RM-02"
  ai_rmf: "GOVERN 1.3, MAP 5.1"
  iso42001: "A.5"
  eu_ai_act: "Art. 9"

checks:
  - id: GV-BR-O-01
    description: "A tier taxonomy exists and defines all four tiers."
    type: min_items
    target: "risk_tiers:4"
    document: organizational-policy
    severity: blocking
    evidence: "Four tier definitions."
  - id: GV-BR-O-02
    description: "Every tier carries an explicit damage cap."
    type: every_has_key
    target: "risk_tiers|damage_cap"
    document: organizational-policy
    severity: blocking
    evidence: "No tier defined without a cap."
  - id: GV-BR-O-03
    description: "Every tier names the controls required at that tier."
    type: every_has_key
    target: "risk_tiers|required_controls"
    document: organizational-policy
    severity: blocking
    evidence: "The link from Charter to the operational matrices."

Cell notes

Charter, Blast radius / Organizational

Structural question. Does the org have a written risk tier taxonomy with explicit damage caps and the matching control requirements per tier?

Owner. AI Governance Council.

Template fragment

§3 (AI Risk Appetite Statement) of ../../templates/organizational-policy.md, specifically the tier table:

Tier	Description	Damage cap	Required controls
1	Read-only diagnostics	None	Charter + Identity + Authorization (read-only)
2	Scoped writes	N records / session, $X / day	Tier 1 + Approval gating + Blast radius
3	Destructive operations	Scope-limited	Tier 2 + Sentinels + Interventions runbooks
4	Production-critical	Per-agent declared	Tier 3 + multi-party approval, off-cluster identity, immutable backups

Audit prompts

- Does the org have a tier taxonomy? Or is "every agent is treated the same"?
- Are damage caps quantitative? "Acceptable risk for innovation" is not a damage cap.
- Is the control matrix per tier enforced at PR review? An agent charter that claims Tier 1 but requests destructive scope must be rejected.

Operational tie-in

The tier table is the input to which Covenants cells must be populated for an agent. The PR-review checklist for new agent charters includes "tier ↔ controls" verification.

Citation

NIST CSF 2.0 GV.RM-01, GV.RM-02 (risk management objectives, risk appetite). NIST AI RMF GOVERN 1.3 (risks identified are managed), MAP 5.1. ISO/IEC 42001 §A.6. EU AI Act Art. 9(3), 9(4) (risk-based approach).

Crosswalk

NIST CSF 2 0	GV.RM-01, GV.RM-02
NIST AI RMF	GOVERN 1.3, MAP 5.1
ISO IEC 42001	§A.6
EU AI ACT	Art. 9(3), Art. 9(4)

Cite this cell: https://agenticcovenants.com/govern/blast-radius/server-side/