June 2026 model refresh (v1.7.0) + dated model fact-sheet with refresh protocol#48
Open
ed-da-goat wants to merge 1 commit into
Open
June 2026 model refresh (v1.7.0) + dated model fact-sheet with refresh protocol#48ed-da-goat wants to merge 1 commit into
ed-da-goat wants to merge 1 commit into
Conversation
Verified against current vendor docs (2026-06-10): - NEW references/models.md: all volatile model facts (IDs, params, behavioral quirks) in one place, each vendor section dated, with a 60-day refresh protocol so the skill re-verifies instead of asserting stale facts. - Claude: updated to the Fable 5 / Opus 4.8 era — adaptive-only thinking, sampling params and prefills removed on 4.7+, effort levels incl. xhigh, Opus 4.8 behavioral shifts (more narration, higher ask-rate, conservative tool/subagent triggering) with corrective prompt snippets. - OpenAI: o3/o4-mini section replaced — the o-series is retired; lineup is GPT-5.5 Instant / GPT-5.4 Thinking / GPT-5.4 Pro. Same reasoning-model rules now route to GPT-5.4 Thinking. - Gemini: current model is 3.1 Pro (3-pro-preview discontinued 2026-03-26); added thinking_level and the missing-info guessing caveat. - DeepSeek: R1 section replaced by V4 (Pro/Flash); legacy deepseek-chat / deepseek-reasoner names retire 2026-07-24; V4 supports tools in thinking. - Kimi K2.5/K2.6 and Qwen 3.5/3.6 facts added to models.md. - Template M generalized from 'Opus 4.7 Task Brief' to current Claude agentic brief; Template E CoT list updated to reasoning-native modes. - patterns.md: nidhinjs#27/nidhinjs#36 updated, new nidhinjs#38 (retired model / dead parameter), fixed corrupted table row at the end of the file.
azagreev
added a commit
to azagreev/prompt-master-za
that referenced
this pull request
Jun 11, 2026
Implements two upstream ideas on our diverged tree. Cortex Code (idea from nidhinjs#15): - add Cortex Code (Snowflake CLI agent) profile to tool-profiles.md - list it in README "Works with" and the tool table Dated model fact-sheet (idea from nidhinjs#48): - add references/models.md — per-vendor volatile facts (IDs, defaults, version-tied params, retirements), each section dated with last-verified and governed by a 60-day re-verify protocol; fast-moving vendors marked "verify" rather than fabricated - add pattern nidhinjs#38 (hardcoded retired model / dead parameter); 37 -> 38 - SKILL.md: re-verify pointer in Tool Routing, "Stale model facts" gotcha, models.md in the Reference Files table - tool-profiles.md header points to models.md as the volatile-facts layer bump 1.9.0 -> 1.10.0 Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
azagreev
added a commit
to azagreev/prompt-master-za
that referenced
this pull request
Jun 22, 2026
….23.0 Generalize the existing `Assumed output format` convention to all settings-as-knobs profiles (Gamma, Perplexity, Grok, image-AI): the skill now surfaces the knob defaults it assumed as an explicit, overridable `Assumed settings:` note line — listing only knobs the user did not set, each with its value + where to change it — without spending a clarifying question. Skipped for prose tools with no knobs (DALL-E 3, Flux). - SKILL.md: Hard rule + Diagnostic + Gotcha preamble generalized (net +0 lines) - tool-profiles.md: Grok / Perplexity / Gamma / Image AI surface-knobs lines - templates.md: Template I / K / N / O Assumed-settings notes - patterns.md: new pattern nidhinjs#48 (count 47 -> 48 across 7 files) - README EN+RU: pattern count 47->48 + rows nidhinjs#47/nidhinjs#48; Routing Index Gamma+ Perplexity rows; profile header "50+ tools across 30+ profiles" - Verified: lint 0/0, opus adversarial gate PASS, clean-room T4 12/12 Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
Model-fact refresh verified against current vendor docs (2026-06-10), plus a structural fix so the skill stops going stale silently.
New:
references/models.mdAll volatile model facts (IDs, params, version-tied behavior) move into one fact-sheet. Each vendor section carries a
last-verifieddate, and the skill is instructed to re-verify any section older than 60 days before asserting it. This addresses the root cause of staleness — facts hardcoded in evergreen instruction text — rather than just this round's symptoms.Fact updates (all verified 2026-06-10)
budget_tokens/sampling params/prefills 400 on 4.7+), effort levels incl.xhigh, and Opus 4.8's behavioral shifts (narrates more than 4.7, asks more often, conservative tool/subagent triggering) with corrective prompt snippets. The old "silent agent → add progress output" advice now backfires on 4.8.gemini-3-pro-previewdiscontinued 2026-03-26); addedthinking_leveland the documented may-guess-when-info-missing caveat.deepseek-chat/deepseek-reasonernames retire 2026-07-24; V4 supports tool calls inside thinking mode.Template/pattern fixes
schema\" |).Why the split matters
v1.6.0's facts were correct at commit time and wrong within a quarter — same will happen to this PR's facts. The dated fact-sheet + refresh protocol makes the skill degrade gracefully (re-verify) instead of confidently asserting last quarter's model landscape.
Happy to split this into smaller PRs if preferred.