PRISM v2.14.0 + Lens Library v0.15 — the output-discipline rev #29
Ronkupper
announced in
Announcements
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
Shipped today, two-track:
What shipped. SP-16 "The Elephant Rule" broadens from negations alone to the uninvited-frame family — negations, defensive intensifiers ("real", "actual"), and temporal hedges ("for now") — with a self-bounding-verb fast path and a worst-case-section exception. Four Standing Principles join the catalog: SP-17 Plain words first, SP-18 It must recompute (arithmetic, structure, citation-claim congruence), SP-19 Claims carry their basis, SP-20 Editions stand alone. The Execution Self-check gains a mechanical step 6 (the SP-18 recompute gate) and its step 5 broadens to the full family. A new Independent Validation Dispatch (§4.14, template Appendix E.13) sends finished deliverables to a fresh session — only the artifact in context, author rationale excluded — for severity-tagged findings across logic, defensibility, consistency, readability, and quality.
Lens track. LL-D-019 "Who said otherwise?" broadens to the family; LL-D-020 "Help or ammunition?" (named-audience per-sentence triple audit) and LL-D-021 "Does a stranger follow?" (cold-reader pass) join Pack 1. Catalog 24 → 26 entries, schema unchanged.
Where it's wired. §10.1.7–§10.1.11 (the SP family), §3.2.2 + Appendix E.2 (Self-check steps 5–6), §4.14 + Appendix E.13 (validation dispatch), §2.1 roster, §2.3/§10.2 tables, §16 calibration item 9, Appendix C.2, glossary. Frontmatter supersedes-pointer corrected (2.12.2 → 2.13.0, a shipped v2.13.0 metadata bug).
Unchanged. Monitor, Probe, and Gate semantics; lint catalog v4 (0 errors); the §2.2 ASCII map.
Report-back (§16 item 9). From real runs: intensifier/hedge false-flag rates, step-6 recompute catches, validation-dispatch yield (findings the producing session had passed), LL-D-020/021 fire rates. These calibrate how strict the family stays and how often the dispatch is worth firing.
Beta Was this translation helpful? Give feedback.
All reactions