| name | 13-ai-agent-slo-doc |
| description | Use when defining AI agent SLOs for task success, safe action completion, intervention, budget, latency, abandonment, and recovery; use ai-hallucination-slo-doc for answer factuality and citation quality. |
| metadata | {"portable":true,"compatible_with":["claude-code","codex"]} |
AI Agent SLO Doc Skill
Use When
- Produce or update service-level objective document from approved project evidence.
- Resolve decisions about SLIs, objectives, error budgets, burn alerts, exclusions, and response policy.
- Prepare a reviewable handoff for Service owners, SRE, and release teams.
Do Not Use When
- The task is primarily owned by monitoring-setup; route there and use this skill only for its named output.
- Required project evidence or decision authority is unavailable and the requester expects a pass, release, certification, or production change.
Required Inputs
| Artefact | Source/provider | Required? | Behaviour when absent |
|---|
| Project _context/, approved requirements, and relevant architecture | Project owner and upstream phase skills | Required | Stop at a gap register; do not invent scope, thresholds, integrations, or owners. |
| Existing artefact, implementation, configuration, and evidence named below | Repository, delivery team, or service owner | Required when updating or assessing | Mark inaccessible items not assessed; do not treat them as passed. |
| Target audience, environment, risk tolerance, and authority | Requester and accountable owner | Required | Produce a read-only outline with explicit assumptions; do not mutate project or production state. |
Outputs
| Artefact | Consumer | Observable acceptance condition |
|---|
| Service-level Objective Document | Service owners, SRE, and release teams | Each SLO has a computable SLI, justified target, data source, exclusions, burn policy, and linked response. |
| Decision and gap register | Reviewer and downstream phase owner | Every assumption, rejected option, unresolved dependency, waiver, and owner is explicit. |
| Validation evidence | Release or governance reviewer | Checks identify command or method, date, result, evidence location, and all unassessed items. |
Evidence Produced
| Evidence | Minimum content | Acceptance |
|---|
| Traceability record | Source artefact, decision, output section, owner | No mandatory decision is source-free. |
| Quality-gate result | Check, expected result, observed result, evidence path | Failures and unavailable checks cannot appear as passes. |
| Review record | Reviewer, date, disposition, open actions | The consumer can reproduce the acceptance decision. |
Capability and Permission Boundaries
- Minimum capabilities: read and search the authorised project sources. Execution is optional and limited to non-destructive validation.
- Assessment and planning default to read-only. Create or edit the named project document only when the request explicitly authorises it. Production mutation, publishing, destructive action, spending, external communication, or certification claims require separate explicit authority.
- Treat secrets, tenant data, incident evidence, and financial records as least-privilege inputs; expose only the minimum evidence needed for review.
Degraded Mode
If files, execution, network, rendering, environment access, fonts, or current evidence are unavailable, return the narrowest useful draft plus a gap register. Label affected checks not assessed, retain the intended acceptance oracle, and state who must supply or verify the missing evidence. Never convert an unavailable check into a pass.
Decision Rules
| Choice | Action | Failure or risk avoided |
|---|
| Evidence is complete and authority is explicit | Choose objectives from user harm and measured baseline and produce the full artefact. | Unmeasurable reliability promises. |
| A required source or approval is missing | Stop the affected branch; record the gap, owner, and unblock condition. | Fabricated requirements or unauthorised action. |
| Evidence conflicts across sources | Preserve both claims, identify the controlling owner, and request a recorded decision. | Silent selection of a convenient but wrong source. |
| A check cannot run in the available environment | Keep its oracle and mark it not assessed; require later execution evidence. | False assurance from capability limits. |
Workflow
- Confirm the named deliverable, consumer, scope, environment, authority, and neighbouring-skill boundary.
- Inventory required sources and validate provenance, freshness, internal consistency, and missing inputs. Stop the affected branch on a mandatory gap.
- Extract traceable requirements, invariants, risks, and measurable acceptance criteria; record conflicts before choosing a design or procedure.
- Apply the decision rules and the domain workflow below. For a failed branch, preserve evidence, choose the documented recovery path, or escalate to the named owner.
- Draft the artefact, decision register, and evidence record together. Do not defer failure handling, rollback, security, tenancy, accessibility, or operational ownership.
- Run available checks, review every result, repair failures, and hand off only when acceptance is observable. If recovery fails or authority is exceeded, stop and escalate without mutation.
Quality Standards
- Ground every section in a named project source, decision, measured result, or accountable owner.
- Give each requirement or procedure a deterministic oracle that another reviewer can reproduce.
- Keep assumptions, exclusions, degraded checks, residual risks, and waivers visible at handoff.
- Preserve the domain invariants and more specific controls in the existing workflow below; this contract does not replace them.
- Run the repository anti-AI-slop gate: remove filler, verify named standards and dependencies, and retain purposeful domain detail.
Anti-Patterns
- Copying a generic template without mapping it to project sources. Fix: attach each section to an approved requirement, configuration, risk, or owner.
- Choosing a threshold because it is common practice. Fix: derive it from a requirement, measured baseline, risk decision, or current verified source.
- Reporting an inaccessible or unexecuted check as passed. Fix: mark it
not assessed, preserve the oracle, and name the verifier.
- Mixing the neighbouring monitoring-setup concern into this artefact without a boundary. Fix: cross-reference its output and keep ownership explicit.
- Omitting failure, rollback, empty-state, security, tenancy, or escalation behaviour. Fix: specify the trigger, safe action, verification, and owner for each applicable case.
- Mutating a repository, environment, tenant, ledger, or external system while drafting guidance. Fix: remain read-only until the exact mutation and authority are explicit.
- Claiming compliance, certification, readiness, or release from prose alone. Fix: require source-attributed evidence and a named acceptance decision.
Worked Example
Given an approved project source and a conflicting implementation detail, record both with provenance, stop the affected branch, and obtain the accountable owner's decision. Then update the relevant contract, define a reproducible acceptance check, and retain its observed result. The artefact is accepted only when each SLO has a computable SLI, justified target, data source, exclusions, burn policy, and linked response.
References
- logic.prompt - load only when its template, logic, or detail is needed.
- README.md - load only when its template, logic, or detail is needed.
Overview
The agent-specific complement to ai-hallucination-slo-doc and the parent SaaS SLO doc. Treats agent task success, intervention rate, irreversible-action incidents, agent-task availability, and agent-cost-per-run as first-class SLIs.
Core Instructions
Step 1: Inventory SLIs per agent feature
Required SLIs:
- Task success SLI — % of agent runs marked successful by the production-sample replay through the judge-LLM (or by user-confirmed completion where available).
- Step efficiency SLI — mean step count of successful runs vs gold-trajectory step count.
- Intervention SLI — % of runs that required mid-run human intervention.
- Irreversible-action-incident SLI — count per million runs of irreversible actions later confirmed incorrect by the user.
- Agent-task availability SLI — % of starts that reach a terminal state within the run's max-wallclock budget (not aborted by infra failure).
- Agent-cost-per-run SLI — mean and P95 USD cost per run, against the per-tenant budget envelope.
- Tool-error rate SLI — % of tool calls that return non-retryable or safety errors.
Step 2: Measurement procedure
For each SLI declare source + sampling:
- Task success: nightly production-sample replayed through the judge-LLM. Sample rate per feature.
- Intervention: emitted by the orchestrator on every run; aggregated per feature.
- Irreversible-action-incident: counted at user-flag time (user marks "this was wrong"); also at admin-review time.
- Availability: emitted by the orchestrator.
- Cost: emitted by the dispatcher (LLM + external API cost rolled up).
Step 3: Per-feature SLO targets by tier
| Tier | Task success | Intervention | Irreversible incidents | Availability | Cost-per-run |
|---|
| Pro | >= 0.90 | <= 20% | 0 (zero budget) | >= 0.99 | within feature cap |
| Enterprise | >= 0.95 | <= 10% | 0 (zero budget) | >= 0.995 | within feature cap |
Step 4: Error budgets
Standard formula: error_budget = (1 - SLO) × runs_in_window. Irreversible-action incidents: zero budget; any breach is SEV1 plus per-tenant kill-switch for the feature.
Step 5: Multi-burn-rate alerts
| Alert | Burn rate | Window | Threshold |
|---|
| Fast burn (task success) | 14x | 1 h | 2% of monthly budget |
| Medium burn | 6x | 6 h | 5% |
| Slow burn | 1x | 3 d | 10% |
| Intervention surge | 3x baseline | 1 h | any feature |
| Irreversible incident | n/a | 0 | any |
| Cost overshoot | per-tenant 200% of envelope | 1 h | throttle then pause |
Step 6: Freeze and pause rules
- Task-success error budget exhausted: freeze planner / catalogue changes; require executive approval for further model bumps.
- Intervention rate up > 50% in 7 d: roll back the last planner/prompt change; SEV2.
- Irreversible-action incident: per-tenant feature kill-switch; SEV1; postmortem; admin notification.
- Cost overshoot: per-tenant throttle then per-tenant pause; SEV2.
Step 7: Customer-facing commitments
Per tier define what is contractually committed. Likely:
- Agent-task availability (numeric).
- Notification of any irreversible-action incident within 24 h.
- A user-facing "this action was performed by an agent" notification standard.
- The user's right to request human handling for any task class.
Step 8: Write the doc
AI_Agent_SLO_Doc.md sections: 1) Agent SLI Inventory, 2) Measurement Procedure, 3) Per-feature SLO Targets, 4) Error Budgets, 5) Burn-Rate Alerts, 6) Freeze & Pause Rules, 7) Customer-Facing Commitments, 8) Review Cadence.
Standards
- Google SRE applied to agents
- ISO/IEC 25010 (functional correctness)
- ISO/IEC 42001 Clause 9
- NIST AI RMF MEASURE
Resources
logic.prompt, README.md, references/ai-agent-slo-template.md.