Creates boundary-point validation contracts, defines invariant-based success criteria, and sets up automated verification probes so reliability workflows trigger on objective evidence rather than intuition. Use when designing robust handoff, memory-persistence, or tool-call reliability workflows; when you need to verify handoffs work, check memory persistence, validate tool calls succeeded, or convert vague reliability goals into concrete, testable checks at each boundary point with explicit failure-class mapping (operational vs. critical); or when you want to test your workflow end-to-end, make sure it works, or verify your automation runs correctly using read-back probes and escalation triggers rather than agent confidence. Includes explicit untrusted-content/prompt-injection guardrails for third-party inputs.
96
Quality
90%
Does it follow best practices?
Impact
98%
1.25xAverage score across 9 eval scenarios
{
"context": "Tests whether the agent correctly handles the memory-resume boundary, including key-exists invariants, timestamp-freshness checks, deserialization validation, and appropriate failure classification and verification probes for memory persistence.",
"type": "weighted_checklist",
"checklist": [
{
"name": "Memory resume boundary",
"description": "contract.md identifies the memory/session resume as a named boundary point",
"max_score": 8
},
{
"name": "Key exists invariant",
"description": "The contract includes an invariant that checks the session key exists in the store (e.g. 'key exists in store', 'store.get(key) returns non-null')",
"max_score": 10
},
{
"name": "Timestamp freshness invariant",
"description": "The contract includes a 'timestamp fresh' or 'timestamp < max_age' invariant for the memory boundary",
"max_score": 10
},
{
"name": "Value deserialises invariant",
"description": "The contract includes an invariant verifying the stored value can be deserialised (e.g. JSON parse, schema check)",
"max_score": 8
},
{
"name": "Table columns present",
"description": "contract.md contains a markdown table with columns: Boundary, Required Invariants, Verification Probes, Failure Class, Escalation Trigger",
"max_score": 8
},
{
"name": "Stale entry as operational",
"description": "The contract classifies a stale/expired entry as 'operational' (not critical) failure class",
"max_score": 10
},
{
"name": "Missing key as critical",
"description": "The contract classifies a missing key as 'critical' failure class",
"max_score": 10
},
{
"name": "Re-computation escalation",
"description": "The contract specifies 'force re-computation' (or equivalent: regenerate context, restart session fresh) as the escalation action",
"max_score": 8
},
{
"name": "Timestamp check in script",
"description": "resume_check.py checks that (current_time - timestamp) is within a max_age threshold (e.g. 300 seconds), using time.time() or datetime",
"max_score": 10
},
{
"name": "Objective checks only",
"description": "Neither contract.md nor resume_check.py uses agent confidence, gut-feel, or qualitative assessment as a trigger — all checks are deterministic comparisons",
"max_score": 10
},
{
"name": "Non-null probe",
"description": "The contract or script includes an explicit check that a store.get or file load returns a non-null/non-empty result",
"max_score": 8
}
]
}