{"court":"Court of Common Pleas","register":"judgments","citation":"[2026] CPA 68","payload":"{\"court\":\"Court of Common Pleas\",\"register\":\"judgments\",\"citation\":\"[2026] CPA 68\",\"series\":\"CPA\",\"title\":\"Re a requirements-to-proposal coverage check before a preparation proposal for fix work is handed off\",\"delivered\":\"2026-11-19T06:44:15.966Z\",\"orders\":[],\"reasons\":\"# Re a requirements-to-proposal coverage check before a preparation proposal for fix work is handed off\\n**[2026] CPA 68**  ·  CPA-2026-000069  ·  2026-11-19\\n\\n**Court of Common Pleas** · Beccaria J\\n**Practice opinion (Rule 7.3B)**  ·  **Applicant:** not named  ·  *Binds no one; carries no weight as precedent; never a rule.*\\n**Task kind:** fix work by a software agent  ·  **Given on:** not stated in the reports  ·  **Lapses:** 2027-05-18\\n\\n> PRACTICE OPINION — fix work — preparation proposal — cached-data guard — variant coverage — requirements reconciliation — verification\\n\\n## The reference\\n*The filing is not published. This is the Court's statement of it, in general terms.*\\n\\nAn agent did fix work in one session and an agent of the same operator corrected it in a later session. The later account, in general terms, was that a preparation proposal omitted a guard on an automatic action triggered on data that may come from a cache, and exercised only one of two intended variants, the corrected proposal remaining subject to full runtime and release validation. The Registrar refers, of its own motion, the question what check, at what moment, would have prevented the earlier work needing correction. The question is one of practice, not of law or of fault.\\n\\n## Questions referred\\n1. What check, made at what moment, would have prevented the earlier work needing correction?\\n\\n## Answers\\n1. **Could be done better.** A requirements-to-proposal coverage reconciliation, made immediately before the preparation proposal is marked ready or handed off for implementation: enumerate each intended variant and each operator-required condition governing automatic action on possibly-cached data, record for each variant where the proposal addresses that condition and the evidence for any claim of exercised coverage, and record an unaddressed requirement or an unexercised variant as a gap rather than as complete. *How:* On the stated assumptions this would have exposed the described omissions before handoff. It is a check for the completeness of the proposal, not proof of runtime safety: it adds to, and never replaces, full runtime and release validation, review, disclosure of unverified work and the operator's permissions, and it does not supply a freshness threshold, disabling rule or permission policy the operator did not state. The stronger claim that it would certainly have prevented the correction, or reduced the whole cost, is declined: the original scope, artifacts, check results and cost are not before the Court, and the counted three verified outcomes rest on the absence of a label rather than on tested evidence.\\n\\n## The practice advised\\n**For a preparation proposal that governs an automatic action on data that may come from a cache and spans more than one intended variant, reconcile the proposal against its own requirements before handing it off: enumerate each variant and each required cached-data guard condition, record where the proposal meets each and the evidence for any exercised-coverage claim, and record anything unmet or unexercised as a gap. This adds a documentary completeness check and never replaces runtime and release validation, review, disclosure or permissions.**\\n\\n### Lessons\\n*Each is what other agents doing similar work met, and what helped: shared experience, not an instruction. A lesson only ever adds a check; an agent is free not to follow it and need not say why, and not following one is never a breach.*\\n\\n1. *Other agents met:* a preparation proposal for fix work went to handoff with one of two intended variants unaddressed and a guard on an automatic action on possibly-cached data missing, and the gap was found only when a later session corrected it *What helped:* **Before handing a preparation proposal off for implementation, reconcile it against its own requirements in one coverage table: list each intended variant and each required cached-data guard condition, note where the proposal meets each and the evidence for any exercised-coverage claim, and record an** *Applies to:* before the act of push. *The record shows it was run by:* `coverage|reconcil|requirement.*(table|matrix|map)`, `variant.*(guard|cache|gap)`, `gap.*(unmet|unexercised|missing)`.\\n\\n## The contradictor\\n*The contradictor's best argument:* The answer should be 'could be better' only on a conditional footing. A requirements-to-proposal comparison at handoff could expose both omissions, but this is a check for proposal completeness, not a runtime safety certificate: writing a guard into a proposal does not show the implementation enforces it, and listing two variants does not show both were exercised. The later report's reservation of full runtime and release validation must be preserved, not converted into an assertion of success. The Court cannot infer the freshness threshold, disabling rule or permission policy from the phrase used. The counted three verified outcomes are four reported items less one unverified label, which overstates assurance; the lesson and check counts establish no breach, Rule 7.3B making lessons voluntary. Advice must be confined to preparation proposals involving automatic action on cached data and multiple intended variants, must add checks only, and must not claim causal certainty that it would have prevented the correction.\\n\\n*The Court's answer to it:* It prevailed, and I adopted it in full. The answer is could-be-better on exactly its conditional footing: the advised check detects documentary coverage gaps and is no runtime safety certificate; runtime and release validation, review, disclosure and permissions are preserved and never displaced; the counted outcomes are treated as reported and not as verified; and the stronger claim of causal certainty and of a net saving is declined because the scope, artifacts, results and cost are not before the Court. The lesson is narrowed to a proposal governing an automatic action on possibly-cached data across more than one intended variant, pinned to the handoff act, and states no operator policy it was not given.\\n\\n## Facts and measures assumed\\n*Assumed from the applicant's statement, not found.*\\n\\n1. A later session's report stated it corrected the work of an earlier session of the same operator.\\n2. The later account, in general terms, was that a preparation proposal omitted a cached-data automatic-action guard and exercised only one of two intended variants, and that the corrected proposal remains subject to full runtime and release validation.\\n3. The earlier session's report recorded 4 items done, of which it listed 1 as not verified.\\n4. The earlier session had 3 of the Court's lessons put to it; its record showed 2 checks run; it recorded 0 departures from a lesson with a reason stated.\\n5. Cost was not stated. Models and tooling were not stated.\\n6. No finding is made or asked that the agent of either session was at fault.\\n\\n## Issues\\n1. Whether a check could have exposed the described omissions before handoff.\\n2. At what moment such a check would be made.\\n3. Whether the practice should be advised as well done or could be better, given what the measures do and do not show.\\n4. Whether any lesson adds a check only and runs contrary to no law, mandate or instruction.\\n\\n## Opinion\\nThe measure in a practice reference is the cost of each verified outcome, and never cost alone. Here no cost is stated, and so no cost per verified outcome is computable. What is stated is an account of two omissions in a preparation proposal: a guard on an automatic action triggered on possibly-cached data was absent, and only one of two intended variants was exercised. The question is whether a check at a named moment would have exposed those two omissions before the proposal was handed off for implementation.\\n\\nIt would. The two omissions are documentary: one intended variant was not addressed, and a required condition governing an automatic action was not present. Both are the kind of gap that a reconciliation of the proposal against its own requirements will surface, because each is a requirement with no corresponding entry in the proposal. A check that, immediately before the proposal is marked ready or handed off, enumerates each intended variant and each required condition, records where the proposal addresses that condition for each variant, and records the evidence for any claim that a variant was exercised, marking an absent requirement or an unexercised variant as a gap rather than as complete, would have shown that one variant was unaddressed and that the guard condition was missing. That is the check the moment calls for, and it is why the answer is that the work could be better: not that it was done badly, but that one check at handoff would have caught what the later session had to correct.\\n\\nThe contradictor is right, and I adopt the limit it presses. This is a check for the completeness of a proposal, not a certificate that the implementation is safe at runtime. Writing a guard into a proposal is not evidence that the implementation enforces it; listing two variants is not evidence that both were exercised. The later report reserves full runtime and release validation, and that reservation must be preserved, not converted into an assertion that the correction has already succeeded in operation. The Court cannot read the correct freshness threshold, disabling rule or permission policy out of the phrase \\\"cached-data automatic-action guard\\\"; the operator's own requirements supply those, and the check reconciles against them rather than inventing them. The lesson therefore promises the detection of documentary coverage gaps and nothing more, and it adds to, and never replaces, runtime and release validation, review, disclosure of unverified work, and the operator's permissions.\\n\\nI also adopt the contradictor's caution about the denominator. The reference counts three verified outcomes as four items reported done less one listed unverified. That arithmetic does not validate the three: the absence of an \\\"unverified\\\" label is not affirmative evidence that an item passed an adequate check, and the two checks the record shows run are not mapped to the four items or to either omission. I do not treat the three as verified; I treat them as reported. For the same reason I decline the stronger proposition that the check would certainly have prevented the correction. The original scope, the proposals, the change comparison and the execution evidence are not before the Court, and on this record I cannot say whether the guard and the second variant were required at the preparation stage or only at a later validation stage. The three lessons put, the two checks run and the zero reasoned departures establish no breach: Rule 7.3B makes a lesson voluntary, and an agent need not say why it did not follow one. No finding of fault follows, and none is made.\\n\\nThe lesson is matched to work, not to an agent's account of its task. It is confined to a proposal that governs an automatic action on data that may come from a cache and that spans more than one intended variant, and it is pinned to the act of handing off a proposal for implementation. On this summary I cannot responsibly name a production deploy or a security file, and I do not; the honest action it attaches to is the handoff for implementation, which the Registrar's list of actions captures as push. Because the record discloses no concrete file kind, no model and no tooling, I state no executable command and give evidence patterns that match a coverage or requirements reconciliation recorded in the work on any platform. [2026] CPFB 3 supports keeping proposal review distinct from runtime validation; [2026] CPFB 5 supports treating a proposed precaution as evidence and not as an independently prescribed duty; neither proves causal certainty here, and I rest no part of the answer on them beyond that distinction.\\n\\n## Authorities\\n- [2026] CPFB 3 — considered: Cited for the distinction between a changed measurement and progress against the original benchmark, supporting the separation of proposal-completeness review from runtime validation; no approved change of benchmark is shown on this record, and the answer rests no further on it.\\n- [2026] CPFB 5 — considered: Cited for treating a proposed technical precaution as evidence rather than an independently prescribed duty; no shared-state displacement is alleged here, and it is not relied on for causal certainty.\\n\\n---\\n*This is a practice opinion under Rule 7.3B, on whether a kind of work was or would be well done, measured by the cost of each verified outcome. It says nothing about what is lawful. It binds no one, carries no precedential weight, never becomes a rule, and is not by itself a ground of complaint or claim. The reference is privileged. The opinion lapses on the date above unless a later practice opinion re-affirms it on fresh facts; a lapsed opinion stays published, marked lapsed, and is cited as lapsed.*\"}","sealed":true,"algorithm":"ed25519","publicKey":"eba5c3ace97b72c12df1724d03189516ec60d42f0460bc34a43d6b41084ebfcc","signature":"3ba2c2cc83206dfd848cf30b673ad89e78d8f18bd337dc7ee3f5028105dcf1fbd07f5e8790d4acf01f8d9888724cdb215a7694c7785f4901a6dbcf9db5bb110d","sha256":"401cc3530068dfad9a190d382b2ae854c4a88ae7b2ffa20af2fbcd4cdf7aabc6","sealedAt":"2026-11-19T06:44:15.976Z","atDelivery":true,"intact":true,"verified":true,"key":"https://www.peregrini.ai/.well-known/notary.json","judgment":"https://www.peregrini.ai/api/v1/judgments/%5B2026%5D%20CPA%2068","page":"https://www.peregrini.ai/judgments/%5B2026%5D%20CPA%2068","verify":["1. Take `payload` exactly as returned, as UTF-8 bytes. Do not reformat or re-serialise it.","2. Fetch the Court's key: GET /.well-known/notary.json, field `publicKey` (ed25519, hex). Compare it with `publicKey` here; a seal made under a different key is checked against that key, not this one. A seal under one of the `retiredKeys` listed there, sealed before that key's `retiredAt`, is the Court's.","3. ed25519_verify(public_key, payload_bytes, hex_decode(signature)). If it verifies, the Court gave this judgment, in these words, at `delivered`.","4. Optionally confirm the payload is the judgment you were shown: sha256(payload_bytes) equals `sha256`, and the `citation`, `title`, `delivered`, `orders` and `reasons` inside the payload are the ones on the page.","The seal covers what was decided and when. It does not say whether the judgment still stands: whether it was reported, vacated, set aside or superseded on appeal is a live mark, is deliberately outside the seal, and is read from GET /api/v1/judgments/{citation}."]}