agentleFS
peterbamuhigirepeterbamuhigire/proposal-skills/skills/ai-agent-proposals/ai-agent-risk-and-responsible-ai/SKILL.md

ai-agent-risk-and-responsible-ai

Skill11 starsChanged 17 days agoMITView on GitHub

Use when drafting an AI-agent risk register or responsible-AI commitment covering autonomous actions, irreversibility, scope, tool injection, identity, kill switches, and contestability.

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Every risk has likelihood, impact, trigger, mitigation, owner, escalation, residual.

    line 93
  2. Distinctive:

    Verify acceptance conditions, permission boundaries, direct references, and anti-slop controls.

    line 169

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

AI-Agent Risk and Responsible AI

Acknowledgement: Shared by Peter Bamuhigire, techguypeter.com, +256 784 464178.

<!-- dual-compat-start -->

Use When

  • The proposal must present an agent risk register inside the Risk Analysis section.
  • The buyer's CISO, DPO, general counsel, ethics committee, or regulator will review the proposal.
  • The bid is in a regulated vertical or an AI-Act-relevant jurisdiction.
  • The agency must demonstrate Responsible-AI agent practice — not promise it.
  • An incumbent or competitor has had a public agentic incident the buyer will reference.

Do Not Use When

  • The engagement is non-agent (use ai-on-saas-risk-and-responsible-ai or risk-management).

Domain Inputs

  • The action catalogue with reversibility classification.
  • The autonomy-level commitment per action class.
  • The oversight model and intervention SLA.
  • The regulator stance and jurisdiction(s).
  • The kill-switch architecture and the named authority.
  • The agent identity and impersonation policy.
  • The sub-processor list (model providers, tool-call APIs).
  • The buyer's escalation expectations.

Domain Method

  1. Populate the Twelve-Entry Agent Risk Register (below). Each entry carries: risk name, description, likelihood, impact, trigger, mitigation, owner, escalation, residual.
  2. Populate the Responsible-AI Agent Commitment section: scope, principles, governance forum, accountable role (Agent Safety Lead), human-final for irreversibility, full audit log, contestability, transparency-to-affected-party, kill-switch drill cadence, sign-off frequency.
  3. Map the commitment to the named regulator(s): AI Act (high-risk articles where applicable), Kenya NCAIS, Nigeria NAIS, South Africa AI policy, Uganda NITA-U guidance, Rwanda AI Policy, sectoral rules.
  4. Pair every risk with a methodology mitigation in ai-agent-methodology and a procurement answer in the agent procurement pack.
  5. Output the Risk Analysis subsection and the Responsible-AI Agent Commitment subsection of the proposal.

The Twelve-Entry Agent Risk Register

#RiskDescriptionPrimary Mitigation
1Autonomy-action incidentAgent acts wrongly inside its autonomy level (wrong refund, wrong escalation, wrong message)Eval thresholds, intervention SLO, drift watch, rollback, prompt regression
2Irreversibility incidentAgent takes an action that cannot be undone (payment, public statement, regulatory filing, deletion)Reversibility classification, L1 gating on irreversible, human-final approval, value bounds, transaction confirmation token
3Accountability disputeA regulator, court, or counterparty seeks to hold a party liable for an agent actionAction audit log, contractual liability allocation, professional indemnity, redress mechanism
4Scope creepAgent's effective scope grows beyond the action catalogue through buyer requests or emergent behaviourAction catalogue gate, change-control clause, monthly catalogue review, anomalous-action alerting
5Multi-agent collusion / failure cascadeAgents in a system reinforce each other's errors or get stuck in loopsOrchestrator-level governance, blast-radius limit, loop detection, replay determinism, anti-collusion guard
6Prompt injection via tool outputAn external system returns content that injects instructions into the agent's contextOutput filtering at tool boundary, system-prompt hardening, content provenance, capability minimisation
7Memory poisoningPersistent memory is corrupted (intentionally or accidentally) and biases future decisionsMemory scoping (per session, per tenant, per task), memory-write review, periodic memory audit, memory reset cadence
8Regulator action on agentic systemRegulator suspends agentic operations, requires conformity assessment, imposes banKill-switch, conformity readiness, sovereign-AI option, jurisdictional fallback, regulator-notification runbook
9Kill-switch failureKill-switch fails in drill or in incidentMulti-layer kill-switch (per-action, per-agent, per-tenant, global), independent authority chain, quarterly drill, post-drill incident review
10Identity / impersonation breachAgent communicates externally without proper identity disclosure, or is mistaken for a human in violation of policyIdentity-and-impersonation policy, signature on every external communication, automated disclosure injection, regulator-aligned transparency
11Escalation overloadIntervention rate exceeds supervisor queue capacity; cases time out; SLA breaches; users harmedIntervention SLO, queue capacity floor, overflow plan, autonomy gating that backs off when queue is saturated
12Legacy-system damageAgent calls a tool that mutates a system in an unintended way (cascading effect, race condition, foreign-key explosion)Tool-call allowlist with parameter bounds, value bounds, transaction wrapper, dry-run, post-call assertion, automatic rollback

Responsible-AI Agent Commitment — Section Structure

  1. Scope — which agent(s) the commitment covers; which action classes; which tenants; which jurisdictions.
  2. Principles — accountability, transparency, fairness, safety, privacy, human oversight, contestability, kill-switch readiness.
  3. Governance forum — Agent Safety Council: who meets, how often, what they decide; agency-side and buyer-side roles.
  4. Accountable role — Agent Safety Lead (named role) accountable for the commitment, with escalation to the SteerCo.
  5. Human-final on irreversibility — every irreversible action requires a named human approver; the agent cannot complete an irreversible action without that approval.
  6. Full audit log — every agent action is logged with the reasoning trace where feasible, the tool call, the parameters, the result, the human decision (if any). Audit completeness SLA ≥ 99 %.
  7. Contestability — any affected user, tenant, or regulator can challenge an agent action through a named channel; SLA for response; a human is the decision-maker on contested actions.
  8. Transparency-to-affected-party — when an agent communicates externally or takes an action that affects a third party, the third party is informed in line with applicable rules (e.g. AI Act Article 50-style transparency).
  9. Kill-switch readiness — kill-switch architecture, named authority, drill cadence (quarterly minimum), post-drill review.
  10. Eval and red-team discipline — golden datasets, thresholds, refresh cadence, regression harness, drift watch.
  11. Incident disclosure — incident classification, notification timing, communications, redress.
  12. Sign-off frequency — quarterly review by the Agent Safety Council; annual sign-off by the SteerCo.

Regulator Mapping (illustrative)

Regulator / standardWhat it requires of agentsHow the commitment addresses it
EU AI Act (high-risk)Conformity assessment, post-market monitoring, transparency, human oversight, accuracy / robustness / cybersecurityAction audit log, eval + red-team, human-final on irreversibility, conformity readiness, transparency-to-affected-party
Kenya NCAISPrinciples-based; accountability, transparency, fairnessNamed accountable role; published model and agent cards; redress channel
Nigeria NAISRisk-based oversight; human accountabilityAutonomy ladder; named authority on irreversible actions; audit log
South Africa AI policy / POPIA interactionAutomated decision-making with significant effects requires interventionHITL on decisioning; contestability channel; data-subject rights honoured
Uganda NITA-U guidanceData protection and digital trustAudit log; sub-processor disclosure; data-residency option
Rwanda AI PolicySovereign-aware AI; transparency; accountabilitySovereign-AI option; transparency-to-affected-party; named accountable role
Sectoral — FS model-risk (SR 11-7-style)Model inventory, independent validation, ongoing monitoringAgent in the model inventory; eval discipline; drift watch; quarterly review
Sectoral — Healthcare adminNon-clinical scope; HITL on clinicalAction catalogue confines agent to admin only; clinical actions excluded
Sectoral — Legal practice rulesLawyer responsibility; client confidentialityHITL on filing / advising; confidentiality bounded by tenant scope

Quality Standards

  • Every risk has likelihood, impact, trigger, mitigation, owner, escalation, residual.
  • Mitigations are anchored in the methodology, not invented for the proposal.
  • Responsible-AI agent commitment is a section, not a sentence; it names an accountable role.
  • The kill-switch drill cadence is set (quarterly minimum) and the first drill is in pilot.
  • Human-final on irreversibility is committed and the named human authority is identified per action class.
  • Transparency-to-affected-party is named, with the rule reference where applicable.
  • Audit-log completeness SLA is a number.
  • The commitment is auditable — the buyer's auditor can verify the audit log, the kill-switch drill log, the contestability log, the red-team scorecard.

Domain Risks

  • "We follow Responsible AI principles" as the only commitment.
  • Risks listed without triggers, owners, or escalation.
  • Mitigations promised but not anchored in the methodology.
  • Irreversibility treated as a UX concern.
  • Kill-switch named but no drill cadence.
  • "Audit log will be available" with no completeness SLA.
  • Contestability deferred to "future work".
  • Agent Safety Lead role unnamed.
  • Multi-agent risk not addressed in multi-agent engagements.

Domain Outputs

  • Agent Risk Register (table form) — drop into proposal risk section or risk-management.
  • Responsible-AI Agent Commitment subsection — drop into proposal 06-methodology close or a standalone section.
  • Regulator Mapping table.
  • Kill-Switch Drill Cadence statement.

Anti-Patterns

  • Inventing a metric, credential, constraint, or buyer position. Fix: cite the supplied source or mark the item as an assumption requiring confirmation.
  • Treating an unavailable check as passed. Fix: mark it not assessed and state the evidence needed to resume.
  • Advancing autonomy without a named gate owner. Fix: require observable evidence, accountable acceptance, and a rollback path.
  • Reusing another sector or use case without reassessment. Fix: retest affected parties, action scope, reversibility, and jurisdiction.
  • Writing acceptance as “satisfactory” or “appropriate”. Fix: define an observable measure, threshold, evidence record, and decision owner.

Inputs

ArtefactSource/providerRequired?Missing-input behaviour
use cases, action catalogue, affected parties, threat evidence, jurisdiction, and risk appetiteBuyer evidence, ToR, approved discovery record, system owner, or measured operating dataYesStop the affected decision; list the missing source and return only a qualified outline or assumption register.

Outputs

ArtefactConsumerAcceptance condition
Agent risk register and responsible-AI commitmentRisk owner, CISO, DPO, legal counsel, and evaluatorScope, assumptions, exclusions, owners, decision logic, and observable acceptance tests are explicit and traceable to supplied evidence.

Evidence Produced

EvidenceConsumerAcceptance condition
agent risk register and responsible-AI commitmentRisk owner, CISO, DPO, legal counsel, and evaluatorEvery load-bearing claim traces to supplied evidence; assumptions, owners, gates, exclusions, and observable acceptance conditions are explicit.

Capability Contract

Default to read-only for discovery, analysis, review, and planning. Minimum capability is access to the supplied artefacts and permission to calculate or inspect evidence. Edit only the requested proposal working copy. Do not change production systems, contact affected parties, publish, spend, certify compliance, or approve autonomous action without explicit authority from the accountable owner.

Degraded Mode

If files, interviews, telemetry, specialist review, network access, or calculation tools are unavailable, produce the narrowest useful qualified result. Mark each unavailable check as not assessed, separate facts from assumptions, lower confidence, and state the evidence needed to resume. An unassessed gate is never a pass.

Decision Rules

ChoiceActionFailure or risk avoided
Accept, mitigate, transfer, or avoidRate inherent risk, name controls and evidence, assign an owner, then assess residual risk against appetite.A generic register that hides autonomous-action exposure.
Required evidence, authority, or accountable owner is missingStop the affected recommendation or commitment and record the gap.Invented evidence or unauthorised autonomy.
Gate evidence is complete and acceptedAdvance only within the approved scope and retain the evidence trace.Scope drift and irreproducible approval.

Workflow

  1. Confirm the consumer, authority, neighbouring-skill route, and required inputs; stop when a mandatory source or accountable owner is missing.
  2. Inspect the evidence and record facts, assumptions, conflicts, and unavailable checks; stop on a failed safety, finance, regulatory, or acceptance gate.
  3. Apply the domain method and decision rules within the qualified scope, retaining an evidence trace.
  4. Draft the contracted output and reconcile it with methodology, work plan, staffing, pricing, risk, and governance; recover by revising the affected scope or control and rerunning the failed gate.
  5. Verify acceptance conditions, permission boundaries, direct references, and anti-slop controls; block release until failed checks are corrected.

Worked Example

An agent can alter a legacy account. Classify the action by reversibility and blast radius, require approval and rollback evidence, name the incident owner, and prohibit agentic mode until the residual risk is accepted.

<!-- dual-compat-end -->

References

  • ai-agent-risk-register-for-proposals — full register template.
  • ai-agent-responsible-ai-commitment — full commitment template.
  • ai-agent-trust-and-compliance-template — trust section linkage.
  • ai-agent-procurement-questionnaire-pack — procurement answers consistent with the register.
  • ai-on-saas-risk-and-responsible-ai — AI-on-SaaS risk register (load alongside for SaaS-embedded agents).
  • risk-management — generic risk register discipline.
  • ai-agent-methodology — methodology that anchors mitigations.
  • ai-agent-compliance-credentials — trust and compliance posture.
  • ai-agent-procurement-and-questionnaire — procurement answers.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.