holycodex 0.16.9-dev.106.1 → 0.16.9-dev.108.1

package/dist/assets/plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "holycodex",
-  "version": "0.16.9-1",
+  "version": "0.16.9-2",
   "description": "Root-directed capabilities and deterministic installed assets for HolyCodex.",
   "author": {
     "name": "David Basile Filho",

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

@@ -5,6 +5,11 @@ description: Use when CI or release gates need follow-through; monitor their ter
 
 # Babysit CI
 
+Root owns orchestration and mutations; Worker.operations observes the assigned
+gates. Dispatch independent checks and review observation concurrently.
+Use only the gates in the user's requested scope. When asked to monitor CI
+only or ignore review bots, omit review observation and review findings.
+
 Discover the repository's development and release topology for pushes and pull
 requests: refs, required checks, reviews, publication mechanism, and gate
 dependencies. Do not assume a provider, branch name, or separate pipeline.
@@ -13,13 +18,15 @@ For a push, bind every observation to the exact pushed ref and SHA. For a pull
 request, bind it to the current head SHA, target branch, pull-request identity,
 and review commit when the provider exposes one. Inspect required checks and
 relevant bot reviews, inline threads, issue comments, and commit comments. A
-new push invalidates all stale CI and review evidence; rediscover the current
-head and repeat validation and review observation for that SHA.
+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.
+Include target-branch or pull-request mergeability evidence when exposed;
+green checks alone do not prove mergeability.
 
 After a development action, observe its exact ref and SHA to terminal state.
 Repair failed checks and actionable review findings, then repeat validation and

package/dist/assets/plugin/skills/babysit-ci/agents/openai.yaml

@@ -1,6 +1,6 @@
 interface:
   display_name: "Babysit CI"
   short_description: "Follow repository CI and release gates to completion."
-  default_prompt: "Use babysit-ci for this CI or release lifecycle."
+  default_prompt: "Use $babysit-ci for this task."
 policy:
   allow_implicit_invocation: true

package/dist/assets/plugin/skills/codebase-onboarding/agents/openai.yaml

@@ -1,6 +1,6 @@
 interface:
   display_name: "Codebase onboarding"
   short_description: "Explain a repository from a bounded read-only map."
-  default_prompt: "Explain how this repository works."
+  default_prompt: "Use $codebase-onboarding for this task."
 policy:
   allow_implicit_invocation: true

package/dist/assets/plugin/skills/commit/agents/openai.yaml

@@ -1,6 +1,6 @@
 interface:
   display_name: "Commit"
   short_description: "Create coherent local commits with exact scope proof."
-  default_prompt: "Open the commit workflow for this invocation."
+  default_prompt: "Use $commit for this task."
 policy:
   allow_implicit_invocation: false

package/dist/assets/plugin/skills/compress/agents/openai.yaml

@@ -1,6 +1,6 @@
 interface:
   display_name: "Compress"
   short_description: "Shorten assigned text without changing its contract or voice."
-  default_prompt: "Open the compression workflow for this invocation."
+  default_prompt: "Use $compress for this task."
 policy:
   allow_implicit_invocation: true

package/dist/assets/plugin/skills/context7-cli/agents/openai.yaml

@@ -1,6 +1,6 @@
 interface:
   display_name: "Context7 CLI"
   short_description: "Find assigned current library and API documentation first."
-  default_prompt: "Open the current documentation lookup workflow for this invocation."
+  default_prompt: "Use $context7-cli for this task."
 policy:
   allow_implicit_invocation: false

package/dist/assets/plugin/skills/continue-work/agents/openai.yaml

@@ -1,6 +1,6 @@
 interface:
   display_name: "Continue work"
   short_description: "Resume the current HolyCodex work from durable state."
-  default_prompt: "Continue the current HolyCodex work."
+  default_prompt: "Use $continue-work for this task."
 policy:
   allow_implicit_invocation: true

package/dist/assets/plugin/skills/debugging/agents/openai.yaml

@@ -1,6 +1,6 @@
 interface:
   display_name: "Debugging"
   short_description: "Diagnose an observable defect through reproduction or strong evidence."
-  default_prompt: "Open the debugging workflow for this invocation."
+  default_prompt: "Use $debugging for this task."
 policy:
   allow_implicit_invocation: false

package/dist/assets/plugin/skills/dev-server/SKILL.md

@@ -0,0 +1,13 @@
+---
+name: dev-server
+description: Use when a shared background development server is needed; let Root start, reuse, and manage one server for the session.
+---
+
+Root only. Reuse a healthy server for the current project; otherwise start the
+repository's development command in the background using the active runtime's
+process controls. Confirm readiness, return its URL, and share the URL with
+specialists so they reuse it. Root retains the process handle. Keep it alive
+across Assignments; restart only when configuration or server state requires
+it. Stop only the process owned by this session when
+no longer needed. Specialists request server changes through their terminal
+report rather than starting competing servers.

package/dist/assets/plugin/skills/dev-server/agents/openai.yaml

@@ -0,0 +1,6 @@
+interface:
+  display_name: "Dev server"
+  short_description: "Root manages one shared background development server."
+  default_prompt: "Use $dev-server for this task."
+policy:
+  allow_implicit_invocation: false

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

@@ -1,6 +1,6 @@
 ---
 name: grill-me
-description: Use when an unspecified choice has multiple materially different outcomes; ask focused questions to resolve the decision before work depends on it.
+description: Use when Root needs an outcome-changing user decision; ask focused questions while independent work continues.
 ---
 
 # Grill me
@@ -10,12 +10,12 @@ behavior, UX, compatibility, dependencies, cost, security, external effects,
 or another meaningful user-visible result. Infer obvious safe defaults and
 continue without questioning routine omissions.
 
-Before asking, finish authorized discovery and reversible preparation that can
-narrow the choice without deciding it. Ask only for the earliest unresolved
-choice required by the next executable phase. After the answer, continue the
-discovery and ask the next dependent question only if it remains material.
-Avoid questionnaires about later phases whose relevance may change.
+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.
 
-Persist the answer through the current Intent or Plan state model. Return to
-Root with the selected decision, its scope, and the evidence or dependency it
-unblocked; use `needs_root_input` when the choice remains unresolved.
+Record the decision in current Intent evidence. Use `needs_root_input` only
+when the missing 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."
-  default_prompt: "Open the material-choice clarification workflow for this invocation."
+  default_prompt: "Use $grill-me for this task."
 policy:
-  allow_implicit_invocation: true
+  allow_implicit_invocation: false

package/dist/assets/plugin/skills/handoff/agents/openai.yaml

@@ -1,6 +1,6 @@
 interface:
   display_name: "Handoff"
   short_description: "Export a redacted projection of current Intent state."
-  default_prompt: "Open the handoff workflow for this invocation."
+  default_prompt: "Use $handoff for this task."
 policy:
   allow_implicit_invocation: false

package/dist/assets/plugin/skills/refactor/agents/openai.yaml

@@ -1,6 +1,6 @@
 interface:
   display_name: "Refactor"
   short_description: "Restructure one decided seam while preserving its contract."
-  default_prompt: "Open the bounded refactor workflow for this invocation."
+  default_prompt: "Use $refactor for this task."
 policy:
   allow_implicit_invocation: true

package/dist/assets/plugin/skills/rules/agents/openai.yaml

@@ -1,6 +1,6 @@
 interface:
   display_name: "Rules"
   short_description: "Trace one repository rule discovery or application failure."
-  default_prompt: "Open the repository rules workflow for this invocation."
+  default_prompt: "Use $rules for this task."
 policy:
   allow_implicit_invocation: false

package/dist/assets/plugin/skills/stop-slop/SKILL.md

@@ -4,6 +4,8 @@ description: Use when assigned prose has predictable AI writing patterns; remove
 ---
 
 Edit only the assigned prose. Remove filler, formulaic structure, vague claims,
-passive voice, and repeated rhythm while preserving meaning and voice. Use the
+unnecessary passive voice, and repeated rhythm while preserving meaning and voice.
+Treat reference examples as candidates, not bans on ordinary grammar. Preserve
+qualifiers, uncertainty, factual distinctions, and technical precision. Use the
 linked [phrase](references/phrases.md), [structure](references/structures.md),
 and [example](references/examples.md) references when needed.

package/dist/assets/plugin/skills/stop-slop/agents/openai.yaml

@@ -1,6 +1,6 @@
 interface:
   display_name: "Stop Slop"
   short_description: "Edit assigned prose to remove predictable AI patterns."
-  default_prompt: "Open the assigned prose editing workflow for this invocation."
+  default_prompt: "Use $stop-slop for this task."
 policy:
   allow_implicit_invocation: true

package/dist/assets/plugin/skills/stop-slop/references/examples.md

@@ -1,69 +1,15 @@
-# Before/After Examples
+# Examples
 
-## Example 1: Throat-Clearing + Binary Contrast
+Before: "It turns out that most teams struggle with alignment. The uncomfortable
+truth is that nobody wants to admit they're confused. And that's okay."
 
-**Before:**
+After: "Most teams struggle with alignment. People may hesitate to admit confusion."
 
-> "Here's the thing: building products is hard. Not because the technology is complex. Because people are complex. Let that sink in."
+Remove the framing while retaining the qualification and uncertainty.
 
-**After:**
+Before: "In today's fast-paced landscape, we need to lean into discomfort and
+navigate uncertainty with clarity. This matters because your competition isn't waiting."
 
-> "Building products is hard. Technology is manageable. People aren't."
+After: "We need to handle uncertainty clearly and keep pace with competitors."
 
-**Changes:** Removed opener, binary contrast structure, and emphasis crutch. Direct statements.
-
----
-
-## Example 2: Filler + Unnecessary Reassurance
-
-**Before:**
-
-> "It turns out that most teams struggle with alignment. The uncomfortable truth is that nobody wants to admit they're confused. And that's okay."
-
-**After:**
-
-> "Teams struggle with alignment. Nobody admits confusion."
-
-**Changes:** Cut hedging ("most"), removed throat-clearing phrases, deleted permission-granting ending.
-
----
-
-## Example 3: Business Jargon Stack
-
-**Before:**
-
-> "In today's fast-paced landscape, we need to lean into discomfort and navigate uncertainty with clarity. This matters because your competition isn't waiting."
-
-**After:**
-
-> "Move faster. Your competition is."
-
-**Changes:** Eliminated jargon entirely. Core message in six words.
-
----
-
-## Example 4: Dramatic Fragmentation
-
-**Before:**
-
-> "Speed. Quality. Cost. You can only pick two. That's it. That's the tradeoff."
-
-**After:**
-
-> "Speed, quality, cost—pick two."
-
-**Changes:** Single sentence. No performative emphasis.
-
----
-
-## Example 5: Rhetorical Setup
-
-**Before:**
-
-> "What if I told you that the best teams don't optimize for productivity? Here's what I mean: they optimize for learning. Think about it."
-
-**After:**
-
-> "The best teams optimize for learning, not productivity."
-
-**Changes:** Direct claim. No rhetorical scaffolding.
+Replace jargon without dropping the original requirement or inventing a new claim.

package/dist/assets/plugin/skills/stop-slop/references/phrases.md

@@ -20,11 +20,11 @@ Remove these announcement phrases. State the content directly.
 - "Here's what I find interesting"
 - "Here's the problem though"
 
-Any "here's what/this/that" construction is throat-clearing before the point. Cut it and state the point.
+Cut an opener when it adds no information.
 
 ## Emphasis Crutches
 
-These add no meaning. Delete them.
+Delete these when they add no meaning.
 
 - "Full stop." / "Period."
 - "Let that sink in."
@@ -52,7 +52,8 @@ Replace with plain language.
 
 ## Adverbs
 
-Kill all adverbs. No -ly words. No softeners, no intensifiers, no hedges.
+Remove empty intensifiers and redundant hedges. Keep adverbs and qualifiers
+that convey degree, timing, uncertainty, or a technical distinction.
 
 Specific offenders:
 

package/dist/assets/plugin/skills/stop-slop/references/structures.md

@@ -1,134 +1,13 @@
-# Structures to Avoid
+# Structures to reconsider
 
-## Binary Contrasts
+Remove formulaic framing when it obscures the content: repeated binary contrasts,
+rhetorical questions immediately answered, manufactured drama, and strings of
+punchy fragments. State the concrete claim and its evidence.
 
-These create false drama. State the point directly.
+Name the actor when it matters. Use active voice when clearer; passive voice is
+useful when the actor is unknown or the object is the focus. Metaphors and ordinary
+sentence openings are acceptable when they serve the assigned voice.
 
-| Pattern                                                       | Problem                        |
-| ------------------------------------------------------------- | ------------------------------ |
-| "Not because X. Because Y." / "Not because X, but because Y." | Telegraphed reversal           |
-| "[X] isn't the problem. [Y] is."                              | Formulaic reframe              |
-| "The answer isn't X. It's Y."                                 | Predictable pivot              |
-| "It feels like X. It's actually Y."                           | Setup/reveal cliche            |
-| "The question isn't X. It's Y."                               | Rhetorical misdirection        |
-| "Not X. But Y." / "not X, it's Y" / "isn't X, it's Y"         | Mechanical contrast            |
-| "It's not this. It's that."                                   | Same formula, different words  |
-| "stops being X and starts being Y"                            | False transformation arc       |
-| "doesn't mean X, but actually Y"                              | Negation-then-assertion crutch |
-| "is about X but not Y"                                        | False distinction              |
-| "not just X but also Y"                                       | Additive hedge                 |
-
-**Instead:** State Y directly. "The problem is Y." "Y matters here." Drop the negation entirely.
-
-## Negative Listing
-
-Listing what something is _not_ before revealing what it _is_. A rhetorical striptease.
-
-| Pattern                               | Problem                           |
-| ------------------------------------- | --------------------------------- |
-| "Not a X... Not a Y... A Z."          | Dramatic buildup through negation |
-| "It wasn't X. It wasn't Y. It was Z." | Same structure, past tense        |
-
-**Instead:** State Z. The reader doesn't need the runway.
-
-## Dramatic Fragmentation
-
-Sentence fragments for emphasis read as manufactured profundity.
-
-| Pattern                                  | Problem                 |
-| ---------------------------------------- | ----------------------- |
-| "[Noun]. That's it. That's the [thing]." | Performative simplicity |
-| "X. And Y. And Z."                       | Staccato drama          |
-| "This unlocks something. [Word]."        | Artificial revelation   |
-
-**Instead:** Complete sentences. Trust content over presentation.
-
-## Rhetorical Setups
-
-These announce insight rather than deliver it.
-
-| Pattern               | Problem                |
-| --------------------- | ---------------------- |
-| "What if [reframe]?"  | Socratic posturing     |
-| "Here's what I mean:" | Redundant preview      |
-| "Think about it:"     | Condescending prompt   |
-| "And that's okay."    | Unnecessary permission |
-
-**Instead:** Make the point. Let readers draw conclusions.
-
-## Formulaic Constructions
-
-| Pattern                   | Problem                     |
-| ------------------------- | --------------------------- |
-| "By the time X, I was Y." | Narrative template          |
-| "X that isn't Y"          | Indirect. Say "X is broken" |
-
-## False Agency
-
-Giving inanimate things human verbs. Complaints don't "become" fixes. Bets don't "live or die." Decisions don't "emerge." A person does something to make those things happen. AI loves this because it avoids naming the actor.
-
-| Pattern                         | Problem                                                           |
-| ------------------------------- | ----------------------------------------------------------------- |
-| "a complaint becomes a fix"     | The complaint did nothing. Someone fixed it.                      |
-| "a bet lives or dies in days"   | Bets don't have lifespans. Someone kills the project or ships it. |
-| "the decision emerges"          | Decisions don't emerge. Someone decides.                          |
-| "the culture shifts"            | Cultures don't shift on their own. People change behavior.        |
-| "the conversation moves toward" | Conversations don't move. Someone steers.                         |
-| "the data tells us"             | Data sits there. Someone reads it and draws a conclusion.         |
-| "the market rewards"            | Markets don't reward. Buyers pay for things.                      |
-
-**Instead:** Name the human. "The team fixed it that week" beats "the complaint becomes a fix." If no specific person fits, use "you" to put the reader in the seat.
-
-## Narrator-from-a-Distance
-
-Floating above the scene instead of putting the reader in it.
-
-| Pattern                   | Problem                 |
-| ------------------------- | ----------------------- |
-| "Nobody designed this."   | Disembodied observation |
-| "This happens because..." | Lecturer voice          |
-| "This is why..."          | Same                    |
-| "People tend to..."       | Armchair sociologist    |
-
-**Instead:** Put the reader in the room. "You don't sit down one day and decide to..." beats "Nobody designed this."
-
-## Passive Voice
-
-Every sentence needs a subject doing something. Passive voice hides the actor and drains energy.
-
-| Pattern                    | Fix                  |
-| -------------------------- | -------------------- |
-| "X was created"            | Name who created it  |
-| "It is believed that"      | Name who believes it |
-| "Mistakes were made"       | Name who made them   |
-| "The decision was reached" | Name who decided     |
-
-**Instead:** Find the actor. Put them at the front of the sentence.
-
-## Sentence Starters to Avoid
-
-| Pattern                                                         | Fix                                             |
-| --------------------------------------------------------------- | ----------------------------------------------- |
-| Sentences starting with What, When, Where, Which, Who, Why, How | Restructure. Lead with the subject or the verb. |
-| Paragraphs starting with "So"                                   | Start with content                              |
-| Sentences starting with "Look,"                                 | Remove                                          |
-
-Wh- openers become a crutch. "What makes this hard is..." becomes "The constraint is..." or better, name the specific constraint.
-
-## Rhythm Patterns
-
-| Pattern                        | Fix                                                 |
-| ------------------------------ | --------------------------------------------------- |
-| Three-item lists               | Use two items or one                                |
-| Questions answered immediately | Let questions breathe or cut them                   |
-| Every paragraph ends punchily  | Vary endings                                        |
-| Em-dashes                      | Remove. Use commas or periods. No em dashes at all. |
-| Staccato fragmentation         | Don't stack short punchy sentences                  |
-| "Not always. Not perfectly."   | Hedging disguised as reassurance                    |
-
-## Word Patterns
-
-| Pattern                                                                                               | Problem                                                    |
-| ----------------------------------------------------------------------------------------------------- | ---------------------------------------------------------- |
-| Lazy extremes (every, always, never, everyone, everybody, nobody)                                     | False authority. Use specifics instead of sweeping claims. |
-| All adverbs (-ly words, "really," "just," "literally," "genuinely," "honestly," "simply," "actually") | Empty emphasis. See phrases.md for full list.              |
+Vary rhythm without arbitrary bans on punctuation, three-item lists, question
+words, or adverbs. Preserve meaningful contrast, qualifications, uncertainty,
+and technical precision. Judge the passage rather than matching forbidden forms.

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

@@ -0,0 +1,17 @@
+---
+name: visual-loop
+description: Use when visual changes need rendered judgment; let Root inspect the current result and assign concrete 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.
+
+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
+states against the user's requirements and repository design system. Report
+the specific blocker if no available method supports the required judgment.
+
+Reuse unaffected evidence. Do not run this loop for logic-only changes.
+Specialists own implementation and delegated interaction checks.

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

@@ -0,0 +1,6 @@
+interface:
+  display_name: "Visual loop"
+  short_description: "Root judges rendered visual changes and assigns repairs."
+  default_prompt: "Use $visual-loop for this task."
+policy:
+  allow_implicit_invocation: false

package/dist/assets/plugin/skills/web-visualize/SKILL.md

@@ -5,10 +5,9 @@ description: Use when the user requests or approves an interactive visualization
 
 Use ChatGPT Sites to create an interactive, user-facing visualization that
 turns the supplied plan, analysis, design, structured information, or other
-useful material into a clear, usable result. When the user explicitly requests
-a visualization, create it without asking again. If you believe one would
-materially improve a task but the user did not request it, ask Root to obtain
-approval before creating it.
+useful material into a clear, usable result. A visualization request authorizes
+its preparation. Root delegates implementation and judges the rendered result
+with visual-loop; follow the active capability's publication boundary.
 
 Return the visualization and enough context to verify that it answers the
 request. If the requested result cannot be completed with an authorized,

package/dist/assets/plugin/skills/web-visualize/agents/openai.yaml

@@ -1,6 +1,6 @@
 interface:
   display_name: "Web visualize"
   short_description: "Turn useful information into an interactive user-facing visualization."
-  default_prompt: "Create the requested interactive visualization."
+  default_prompt: "Use $web-visualize for this task."
 policy:
   allow_implicit_invocation: true

package/dist/assets/plugin/skills/writing-instructions/SKILL.md

@@ -5,65 +5,33 @@ description: Use when model-facing instructions need authorship or review; make
 
 # Writing instructions
 
-Use this skill for model-facing instructions, including Root/session policy,
-specialist and Role.task contracts, skills, and conditional workflows. Keep
-shared behavior under one owner across model routes; account for model-specific
-differences only when supported by evidence.
-
-Before writing, identify the receiver's effective context: higher-priority and
-user instructions, repository instructions, Role/task policy, relevant skills,
-tools and capabilities, configuration, and current task context. Add only the
-missing semantic delta. Resolve conflicts against the governing authority;
-do not duplicate a meaning already supplied to the same receiver.
-
-Give each meaning one authoritative owner at the narrowest scope that applies:
-Root/session invariants in Root policy, repository conventions in AGENTS.md,
-shared specialist boundaries in the specialist baseline, route-specific behavior
-in Role.task, and genuinely conditional workflow or tool knowledge in a skill.
-Generated projections derive from or validate against that owner. Invocation
-metadata selects a branch; it does not become a second policy source.
-
-Write instructions as short as possible without losing information that
-materially constrains correct behavior. Preserve known task facts, defects,
-distinctions, invariants, authority boundaries, required outcomes, and
-acceptance conditions; do not make the receiver rediscover relevant known
-facts. For implementation details that are not part of the contract and can be
-reliably discovered during execution, point toward the authoritative source
-instead of embedding them. Prefer generic guidance where the meaning is
-general. Name exact implementation details only when needed to locate
-authority, describe a known defect, preserve a distinction, or constrain exact
-behavior.
-
-Define the required outcome and completion evidence, bounded authority, relevant
-constraints, material decisions or blockers that escalate, and stopping
-conditions. Keep required lifecycle and safety invariants explicit. Give the
-receiver room to choose routine safe, reversible, in-scope steps. A contract must
-carry authorized work through its requested terminal state, including required
-repair and proof; a first pass or pending gate is not completion.
-
-Remove older-model competence scaffolding, fixed recipes without a required
-ordering constraint, broad reading rituals, stale environment facts, duplicate
-policy, and unnecessary questionnaires. Preserve meaningful repository checks
-and review gates through their canonical owners instead of reproducing them in
-every skill. Use configuration for verbosity rather than generic style padding.
-
-Use a skill for a distinct conditional workflow or tool capability. Keep its
-description short and specific enough to distinguish the invocation from
-nearby skills. Avoid broad triggers that compete for ordinary work; audit
-implicit invocation against the other skills the receiver can load.
-
-Keep the main body focused on the shared outcome, boundaries, and acceptance
-criteria. Link branch-specific tool knowledge or uncommon procedures from the
-condition that needs them, so only the selected branch loads that detail. Do
-not split a short coherent rule merely to create more files.
-
-Keep display and invocation metadata in agents/openai.yaml and behavior in its
-single authoritative source. References should supply a missing decision or
-mechanic, not copies of the body or global policy. Preserve provenance and
-license attribution under the original upstream names when renaming a skill.
-
-An instruction change is complete when the receiver can act within its authority,
-recognize completion and escalation boundaries, and return sufficient evidence;
-the effective context has no contradictory or duplicate rule for that meaning.
-Validate changed behavior with meaningful proof and repository-required gates,
-using the canonical testing policy rather than inventing extra test rituals.
+Identify the receiver's effective context: instruction hierarchy, repository
+rules, Root or Role.task policy, selected skills, tools, configuration, and
+Assignment facts. Add only the missing constraints; resolve contradictions
+against the governing authority. User instructions override skill guidelines.
+
+Give each meaning one owner: Root/session policy for orchestration and judgment,
+AGENTS.md for repository conventions, the specialist baseline for shared
+boundaries, Role.task for task authority, and skills for conditional workflows
+or tool knowledge. Generated projections derive from that owner. Metadata
+selects the workflow; it does not duplicate its policy.
+
+State the outcome, bounded authority, material constraints, completion evidence,
+and escalation conditions. Preserve known facts needed to act; point to sources
+for details the receiver can discover reliably. Routine safe in-scope choices
+should proceed without questions. Carry authorized work through repair and proof.
+
+Remove redundant policy, older-model scaffolding, broad reading rituals, rigid
+recipes without a real dependency, hypothetical gates, and repeated testing.
+Preserve lifecycle invariants and required checks. Use configuration for default
+verbosity; use prompts for task-specific evidence and output requirements.
+
+Keep skill descriptions short and situation-first. Audit triggers alongside
+nearby skills; avoid skills that merely restate a task contract. Keep coherent
+shared guidance together and link uncommon branch-specific mechanics only from
+the condition that needs them. Metadata default prompts invoke the named skill
+without embedding another instruction source. Preserve upstream attribution.
+
+Validate the effective projection, not exact prose: the receiver can act within
+its authority, recognize completion and blockers, and return sufficient evidence
+without contradictory rules. Reuse the canonical testing policy.

package/dist/assets/plugin/skills/writing-instructions/agents/openai.yaml

@@ -1,6 +1,6 @@
 interface:
   display_name: "Writing instructions"
   short_description: "Author and review model-facing instructions."
-  default_prompt: "Use writing-instructions for this instruction change."
+  default_prompt: "Use $writing-instructions for this task."
 policy:
   allow_implicit_invocation: true