agent-loop-testing

Agent Loop Testing

When to Use

Use this skill when you need evidence-bounded analysis, design, or validation preparation for loop state, plan/action/observation cycles, stop conditions, budgets, repetition, and trace evidence.
Use it to review an Agent, RAG, or LLM plan, result, or evidence set and produce actionable improvements.
Use it when context is incomplete but a bounded first pass with assumptions, gaps, and human-decision boundaries is still useful.

Default to Markdown organized by domain risk, evidence state, priority, and boundary.
When the user requests tables, CSV, JSON, or ticket fields, preserve the same finding fields, evidence, and decision boundaries.
Before machine consumption, confirm the schema, enums, required fields, and evidence sources.

Read and follow prompts/agent-loop-testing.md, including its input audit, domain coverage, and output order.
Extract scope, environment, version, time window, constraints, success criteria, and available evidence, with attention to loop state, plan/action/observation, stop condition, budget, repetition.
Separate confirmed facts, evidence-backed inferences, candidate recommendations, and Human decisions before ranking by risk and evidence strength.
Turn high-risk items into preconditions, steps, expected behavior or decision criteria, required evidence, and a validation method.
When input is incomplete, deliver a bounded first pass, state unsupported conclusions, and never present static material as execution evidence.

When calibrating domain outcomes, valid controls, or evidence gaps, read references/domain-decision-example.md for sanitized input, per-item oracles, and reproducible calculations.
Always read prompts/agent-loop-testing.md; it is the complete execution specification for this skill.
For evaluation, read evals/eval.yaml and the matching cases under evals/cases/.
Load references/, examples/, scripts/, or output-formats.md only when those directories exist and the task needs them.

Record task/run/step IDs, plan, tool and normalized arguments, observation identity/digest, business state delta, and cumulative time/cost; redact sensitive values while preserving deduplication evidence.
Progress requires new goal-relevant evidence or state advancement. Repeated action/observation with no advancement counts as a no-progress cycle; successful calls or repeated prose do not establish progress.
Apply supplied no-progress limits, retry policy, and time/cost budgets. Safety/cancellation stops immediately; any exhausted hard budget precedes more planning. Ask for missing limits rather than invent numbers.
Repetition takes the project-authorized change-strategy, stop, or escalation branch. Report trigger step, termination class, remaining work, and absence of extra effects.

Keep the analysis focused on loop state, plan/action/observation cycles, stop conditions, budgets, repetition, and trace evidence; do not replace business owners or Human risk acceptance, exception approval, or safety decisions.
Never invent system behavior, fields, model outputs, data, thresholds, root causes, execution records, or pass claims.
Static design, plans, file presence, or a dry run retain their evidence state and cannot become proof of real execution.
When evidence is insufficient, use pending confirmation, blocked, unassessed, or NOT_SCORED and give the smallest validation method.
For user data, production, or safety work, use least privilege, masked data, mocks, dry runs, or isolation.

Covered loop state, plan/action/observation, stop condition, budget, repetition, with source, evidence state, and validation method for each.
Separated facts, inferences, candidate recommendations, gaps, and Human decisions.
Gave high-risk items P0/P1/P2/P3 or an equivalent priority, owner role, and close condition.
Did not turn plans, static checks, or dry runs into test execution, all-passed, or safety-approved claims.
Stated residual risk, stop/escalation conditions, and next actions.

Listing checks without triggers, expected concerns, owner roles, close conditions, and evidence.
Treating adjacent tests or model tools as a complete Agent loop judgment.
Using unexplained numbers for false precision or writing correlation as causation.
Refusing incomplete input, or pretending that incomplete evidence is conclusive.

Start with paths most likely to cause user harm, business loss, or decision blockage.
Use the smallest verifiable experiment to reduce uncertainty and record conditions, versions, sources, and evidence.
Make the Skill independently installable, executable, and reviewable by another engineer.