Skip to content

v5.28.0 — Rationalization Tail: De-Theater & Gate-Keeper Unification

Choose a tag to compare

View all tags
@samibs samibs released this 13 Jul 09:21
· 18 commits to main since this release
v5.28.0
0f7634d
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

The finishing pass on the skill rationalization (v5.27.0).

One gate-keeper

The framework had two gate keepers — /gate-keeper (quality + an evolved capability-gate model) and the standalone reptilian-gate-keeper (the original NASAB-Pillar-3 agent, whose function /gate-keeper had already absorbed). The reptilian one was referenced by nothing, so it was retired (96 skills); its unique four-dimension Evidence Collection Protocol was grafted into /gate-keeper.

De-theater

The persona framing was written to strong-arm weaker models — "ruthless senior engineer", "the Merciless Evaluator", "cold-blooded architect", NASAB "Reptilian Gates", crocodile/apex-predator metaphors. A capable model doesn't need theater to be exacting. So the coercion language became plain professional tone that keeps every bit of the rigor:

  • "ruthless / merciless / cold-blooded" → "rigorous / exacting / unsparingly honest"
  • "Reptilian Gates" → "Capability Gates"; animal stages (Hatchling→Apex) → plain (Syntax→Autonomy)

Deliberately kept: brand names (Forge/Anvil/Specter/ColdStart), the cold-blooded tone-config value (renaming breaks existing config.json), agent command ids, and the Zero Tolerance Banned Patterns policy. No renames, no capability change.

Full detail in the CHANGELOG and release notes. Full suite: 198 passing / 0 failing.

Assets 2
Loading