holycodex 0.16.9-dev.108.1 → 0.16.9-dev.110.1

package/dist/assets/plugin/skills/babysit-ci/SKILL.md

@@ -21,10 +21,12 @@ relevant bot reviews, inline threads, issue comments, and commit comments. A
 new push requires CI and review evidence for the new head; reuse unchanged
 topology and local proof that remains applicable.
 
-Use bounded observer waits. Distinguish a bot that is absent or not configured
-from a known bot with no terminal signal, and both from a clean terminal
-disposition. Absence or no signal is an evidence gap. Green checks do not
-resolve outstanding review findings or threads.
+Use bounded observer waits at the longest supported event interval. If the provider
+offers only snapshots, avoid repeated reads before state can change; stop as
+soon as the exact target reaches a terminal result. Distinguish a bot that is absent
+or not configured from a known bot with no terminal signal, and both from a clean
+terminal disposition. Absence or no signal is an evidence gap.
+Green checks do not resolve outstanding review findings or threads.
 Include target-branch or pull-request mergeability evidence when exposed;
 green checks alone do not prove mergeability.
 

package/dist/assets/plugin/skills/grill-me/SKILL.md

@@ -1,21 +1,21 @@
 ---
 name: grill-me
-description: Use when Root needs an outcome-changing user decision; ask focused questions while independent work continues.
+description: Use when a material user decision blocks authorized work; ask Root's concise question through request_user_input.
 ---
 
 # Grill me
 
-Use this skill only for a material unresolved choice involving architecture,
-behavior, UX, compatibility, dependencies, cost, security, external effects,
-or another meaningful user-visible result. Infer obvious safe defaults and
-continue without questioning routine omissions.
+Use this skill only when an unresolved material user decision blocks the next
+step. It is not for routine omissions, safe implementation choices, or
+planning.
 
-Root only. Ask the smallest set of questions needed for the blocked decision,
-in waves and in the user's language, using existing evidence and authorization.
-Never ask what is already known. If no material questions remain, orchestrate
-immediately. Continue independent authorized
-work while awaiting the answer. Do not delay a necessary question for a
-discovery ritual or ask about hypothetical later phases.
+Root only. For every clarification, call `request_user_input` with the smallest
+set of questions needed to resolve the decision, in the user's language and
+grounded in known evidence. Never ask what the session already establishes.
+If no material question remains, continue without invoking this skill.
 
-Record the decision in current Intent evidence. Use `needs_root_input` only
-when the missing answer prevents further progress.
+Continue independent authorized work while awaiting the answer. Do not ask
+about hypothetical later phases. Specialists must not ask the user; they
+return a `needs_root_input` outcome to Root with the specific decision and
+why it blocks their Assignment. Root uses that outcome only when the answer
+prevents further progress.

package/dist/assets/plugin/skills/grill-me/agents/openai.yaml

@@ -1,6 +1,6 @@
 interface:
   display_name: "Grill me"
-  short_description: "Ask only material, dependency-ordered clarification questions."
+  short_description: "Ask Root-only material questions through request_user_input."
   default_prompt: "Use $grill-me for this task."
 policy:
   allow_implicit_invocation: false

package/dist/assets/plugin/skills/visual-loop/SKILL.md

@@ -1,12 +1,16 @@
 ---
 name: visual-loop
-description: Use when visual changes need rendered judgment; let Root inspect the current result and assign concrete repairs.
+description: Use when visual changes need rendered judgment; Root judges the result after specialists inspect their own work and assigns repairs.
 ---
 
 Root only, for visual tasks. Orchestrate implementation, inspect its rendered
 result, judge the visual output, dispatch targeted changes to specialists, and
 repeat until the result is excellent and meets the user's requirements.
 
+Specialists implementing visual changes inspect and judge their own current
+render, repair visible defects, and recheck before returning. This local
+implementation loop supplements Root's higher-level visual acceptance.
+
 Follow the visual-inspection order generated for Root's selected capabilities.
 Unselected capabilities are not fallbacks. Honor higher-priority active-surface
 tool rules. Use the shared development server when needed. Inspect relevant viewports and

package/dist/assets/plugin/skills/visual-loop/agents/openai.yaml

@@ -1,6 +1,6 @@
 interface:
   display_name: "Visual loop"
-  short_description: "Root judges rendered visual changes and assigns repairs."
+  short_description: "Root judges rendered changes after specialists inspect their work and assigns repairs."
   default_prompt: "Use $visual-loop for this task."
 policy:
   allow_implicit_invocation: false