@pigcloud/skills 1.0.10 → 1.1.2

package/skills/05-delivery/release-validation/SKILL.md

@@ -0,0 +1,29 @@
+---
+name: release-validation
+description: Validate the release outcome after deployment or merge.
+---
+
+# Release Validation
+
+## Purpose
+
+Check that the delivered change behaves correctly in the target environment.
+
+## When to Use
+
+- You need post-release verification.
+- The deployment or merge has already happened.
+- You need to compare the real environment against the expected result.
+
+## How
+
+- Confirm the delivered behavior and release status after deployment is complete.
+- Run the smallest useful smoke checks against the real target environment.
+- Capture the actual outcome, any mismatch, and the evidence source.
+- Note whether the release is healthy, degraded, or needs rollback/mitigation.
+
+## When to Avoid
+
+- The delivery is not ready yet.
+- You need to change code rather than document release steps.
+- The request is only a short change summary.

package/skills/05-delivery/release-validation/assets/golden-prompt-suite.current.js

@@ -0,0 +1,22 @@
+#!/usr/bin/env node
+
+module.exports = {
+  "name": "release-validation.golden-prompt-suite.current",
+  "skill": "release-validation",
+  "description": "Validate the release outcome after deployment or merge.",
+  "cases": [
+    {
+      "id": "release-post-check",
+      "prompt": "After release, confirm the core behavior and release result are healthy.",
+      "expectedFirstSkill": "release-validation",
+      "expectedOutput": [
+        "verification result",
+        "release status",
+        "issues"
+      ],
+      "expectedStop": "Do not redesign the release plan.",
+      "expectedHandoff": "change-log"
+    }
+  ]
+};
+

package/skills/05-delivery/release-validation/references/README.md

@@ -0,0 +1,17 @@
+# release-validation References
+
+This directory is local to the skill and is the only reference entry point for this skill.
+
+## Contents
+
+- `../skill-spec.json`: design source for regenerating this skill
+- `prompt-template.md`: reusable prompt wording for this skill
+- `cases.md`: human-readable replay case for this skill
+- `../assets/golden-prompt-suite.current.js`: machine-readable local case fixture
+
+## Guidance
+
+- Keep skill-specific explanation here instead of pushing it into a stage-level shared library.
+- Keep shared hard constraints in `../../../rules/index.md`.
+- Keep canonical names unchanged.
+- Skill summary: Validate the release outcome after deployment or merge.

package/skills/05-delivery/release-validation/references/cases.md

@@ -0,0 +1,26 @@
+# release-validation
+
+## Replay Case
+
+- Case: `release-post-check`
+- First skill: `release-validation`
+
+### Prompt
+
+```text
+After release, confirm the core behavior and release result are healthy.
+```
+
+### Expected Output
+
+- verification result
+- release status
+- issues
+
+### Stop Rule
+
+- Do not redesign the release plan.
+
+### Handoff
+
+- change-log

package/skills/05-delivery/release-validation/references/prompt-template.md

@@ -0,0 +1,18 @@
+# release-validation
+
+## Prompt Template
+
+```text
+You are handling the release-validation skill.
+Goal: Validate the release outcome after deployment or merge
+Inputs: the request, supporting evidence, constraints, and existing repository facts.
+Output contract: verification result, release status, issues.
+Stop rule: Do not redesign the release plan.
+Handoff: change-log.
+Keep the response scoped to this stage and do not absorb the work of the next stage.
+```
+
+## Notes
+
+- Keep the template aligned with the local replay case and local assets.
+- Read `cases.md` for the matching replay example.

package/skills/05-delivery/release-validation/skill-spec.json

@@ -0,0 +1,25 @@
+{
+  "group": "05-delivery",
+  "name": "release-validation",
+  "description": "Validate the release outcome after deployment or merge.",
+  "title": "Release Validation",
+  "purpose": "Check that the delivered change behaves correctly in the target environment.",
+  "whenToUse": [
+    "You need post-release verification.",
+    "The deployment or merge has already happened.",
+    "You need to compare the real environment against the expected result."
+  ],
+  "how": [
+    "Confirm the delivered behavior and release status after deployment is complete.",
+    "Run the smallest useful smoke checks against the real target environment.",
+    "Capture the actual outcome, any mismatch, and the evidence source.",
+    "Note whether the release is healthy, degraded, or needs rollback/mitigation."
+  ],
+  "whenToAvoid": [
+    "The delivery is not ready yet.",
+    "You need to change code rather than document release steps.",
+    "The request is only a short change summary."
+  ],
+  "caseId": "release-post-check",
+  "source": "bootstrap-from-skill-md"
+}

package/skills/shared/codebase-learning/SKILL.md

@@ -0,0 +1,29 @@
+---
+name: codebase-learning
+description: Learn the repository structure, conventions, and current implementation facts.
+---
+
+# Codebase Learning
+
+## Purpose
+
+Build accurate context from the codebase before making decisions or changes.
+
+## When to Use
+
+- You are entering an unfamiliar repository or module.
+- You need to gather facts before proposing a change.
+- The current implementation is the source of truth for the next decision.
+
+## How
+
+- Read the repo structure and key files first, then summarize the stable facts that matter.
+- Focus on durable facts: directory layout, entry points, conventions, and live behavior.
+- Separate observations from inferences so downstream work can trust the summary.
+- Capture only the facts that a future task is likely to reuse.
+
+## When to Avoid
+
+- The relevant facts are already known.
+- You need implementation or review instead of exploration.
+- The task is not about building initial repo context.

package/skills/shared/codebase-learning/assets/golden-prompt-suite.current.js

@@ -0,0 +1,22 @@
+#!/usr/bin/env node
+
+module.exports = {
+  "name": "codebase-learning.golden-prompt-suite.current",
+  "skill": "codebase-learning",
+  "description": "Learn the repository structure, conventions, and current implementation facts.",
+  "cases": [
+    {
+      "id": "repo-orientation",
+      "prompt": "I just entered this repository. Help me quickly map the structure, conventions, and key facts.",
+      "expectedFirstSkill": "codebase-learning",
+      "expectedOutput": [
+        "repo map",
+        "conventions",
+        "key facts"
+      ],
+      "expectedStop": "Do not propose implementation changes yet.",
+      "expectedHandoff": "evidence-collector"
+    }
+  ]
+};
+

package/skills/shared/codebase-learning/references/README.md

@@ -0,0 +1,17 @@
+# codebase-learning References
+
+This directory is local to the skill and is the only reference entry point for this skill.
+
+## Contents
+
+- `../skill-spec.json`: design source for regenerating this skill
+- `prompt-template.md`: reusable prompt wording for this skill
+- `cases.md`: human-readable replay case for this skill
+- `../assets/golden-prompt-suite.current.js`: machine-readable local case fixture
+
+## Guidance
+
+- Keep skill-specific explanation here instead of pushing it into a stage-level shared library.
+- Keep shared hard constraints in `../../../rules/index.md`.
+- Keep canonical names unchanged.
+- Skill summary: Learn the repository structure, conventions, and current implementation facts.

package/skills/shared/codebase-learning/references/cases.md

@@ -0,0 +1,26 @@
+# codebase-learning
+
+## Replay Case
+
+- Case: `repo-orientation`
+- First skill: `codebase-learning`
+
+### Prompt
+
+```text
+I just entered this repository. Help me quickly map the structure, conventions, and key facts.
+```
+
+### Expected Output
+
+- repo map
+- conventions
+- key facts
+
+### Stop Rule
+
+- Do not propose implementation changes yet.
+
+### Handoff
+
+- evidence-collector

package/skills/shared/codebase-learning/references/prompt-template.md

@@ -0,0 +1,18 @@
+# codebase-learning
+
+## Prompt Template
+
+```text
+You are handling the codebase-learning skill.
+Goal: Learn the repository structure, conventions, and current implementation facts
+Inputs: the request, supporting evidence, constraints, and existing repository facts.
+Output contract: repo map, conventions, key facts.
+Stop rule: Do not propose implementation changes yet.
+Handoff: evidence-collector.
+Keep the response scoped to this stage and do not absorb the work of the next stage.
+```
+
+## Notes
+
+- Keep the template aligned with the local replay case and local assets.
+- Read `cases.md` for the matching replay example.

package/skills/shared/codebase-learning/skill-spec.json

@@ -0,0 +1,25 @@
+{
+  "group": "shared",
+  "name": "codebase-learning",
+  "description": "Learn the repository structure, conventions, and current implementation facts.",
+  "title": "Codebase Learning",
+  "purpose": "Build accurate context from the codebase before making decisions or changes.",
+  "whenToUse": [
+    "You are entering an unfamiliar repository or module.",
+    "You need to gather facts before proposing a change.",
+    "The current implementation is the source of truth for the next decision."
+  ],
+  "how": [
+    "Read the repo structure and key files first, then summarize the stable facts that matter.",
+    "Focus on durable facts: directory layout, entry points, conventions, and live behavior.",
+    "Separate observations from inferences so downstream work can trust the summary.",
+    "Capture only the facts that a future task is likely to reuse."
+  ],
+  "whenToAvoid": [
+    "The relevant facts are already known.",
+    "You need implementation or review instead of exploration.",
+    "The task is not about building initial repo context."
+  ],
+  "caseId": "repo-orientation",
+  "source": "bootstrap-from-skill-md"
+}

package/skills/shared/evidence-collector/SKILL.md

@@ -0,0 +1,29 @@
+---
+name: evidence-collector
+description: Collect code, test, log, and document evidence for conclusions and handoff.
+---
+
+# Evidence Collector
+
+## Purpose
+
+Gather verifiable evidence that supports a decision, finding, or handoff.
+
+## When to Use
+
+- You need proof for a recommendation or conclusion.
+- The work should be backed by traceable sources.
+- A downstream reader needs to trust the exact source of a claim.
+
+## How
+
+- Gather concrete facts, preserve their sources, and package them for downstream reuse.
+- Note what each fact proves and where any inference begins.
+- Keep the evidence small, specific, and easy to re-check.
+- Prefer direct sources such as code, logs, tests, or docs over summaries.
+
+## When to Avoid
+
+- You only need to explore or plan.
+- There is no concrete evidence to collect yet.
+- The task is implementation rather than proof gathering.

package/skills/shared/evidence-collector/assets/golden-prompt-suite.current.js

@@ -0,0 +1,22 @@
+#!/usr/bin/env node
+
+module.exports = {
+  "name": "evidence-collector.golden-prompt-suite.current",
+  "skill": "evidence-collector",
+  "description": "Collect code, test, log, and document evidence for conclusions and handoff.",
+  "cases": [
+    {
+      "id": "evidence-pack",
+      "prompt": "Organize the current evidence into a reusable fact pack.",
+      "expectedFirstSkill": "evidence-collector",
+      "expectedOutput": [
+        "evidence",
+        "facts",
+        "handoff"
+      ],
+      "expectedStop": "Do not reopen the design debate.",
+      "expectedHandoff": "framework-guide"
+    }
+  ]
+};
+

package/skills/shared/evidence-collector/references/README.md

@@ -0,0 +1,17 @@
+# evidence-collector References
+
+This directory is local to the skill and is the only reference entry point for this skill.
+
+## Contents
+
+- `../skill-spec.json`: design source for regenerating this skill
+- `prompt-template.md`: reusable prompt wording for this skill
+- `cases.md`: human-readable replay case for this skill
+- `../assets/golden-prompt-suite.current.js`: machine-readable local case fixture
+
+## Guidance
+
+- Keep skill-specific explanation here instead of pushing it into a stage-level shared library.
+- Keep shared hard constraints in `../../../rules/index.md`.
+- Keep canonical names unchanged.
+- Skill summary: Collect code, test, log, and document evidence for conclusions and handoff.

package/skills/shared/evidence-collector/references/cases.md

@@ -0,0 +1,26 @@
+# evidence-collector
+
+## Replay Case
+
+- Case: `evidence-pack`
+- First skill: `evidence-collector`
+
+### Prompt
+
+```text
+Organize the current evidence into a reusable fact pack.
+```
+
+### Expected Output
+
+- evidence
+- facts
+- handoff
+
+### Stop Rule
+
+- Do not reopen the design debate.
+
+### Handoff
+
+- framework-guide

package/skills/shared/evidence-collector/references/prompt-template.md

@@ -0,0 +1,18 @@
+# evidence-collector
+
+## Prompt Template
+
+```text
+You are handling the evidence-collector skill.
+Goal: Collect code, test, log, and document evidence for conclusions and handoff
+Inputs: the request, supporting evidence, constraints, and existing repository facts.
+Output contract: evidence, facts, handoff.
+Stop rule: Do not reopen the design debate.
+Handoff: framework-guide.
+Keep the response scoped to this stage and do not absorb the work of the next stage.
+```
+
+## Notes
+
+- Keep the template aligned with the local replay case and local assets.
+- Read `cases.md` for the matching replay example.

package/skills/shared/evidence-collector/skill-spec.json

@@ -0,0 +1,25 @@
+{
+  "group": "shared",
+  "name": "evidence-collector",
+  "description": "Collect code, test, log, and document evidence for conclusions and handoff.",
+  "title": "Evidence Collector",
+  "purpose": "Gather verifiable evidence that supports a decision, finding, or handoff.",
+  "whenToUse": [
+    "You need proof for a recommendation or conclusion.",
+    "The work should be backed by traceable sources.",
+    "A downstream reader needs to trust the exact source of a claim."
+  ],
+  "how": [
+    "Gather concrete facts, preserve their sources, and package them for downstream reuse.",
+    "Note what each fact proves and where any inference begins.",
+    "Keep the evidence small, specific, and easy to re-check.",
+    "Prefer direct sources such as code, logs, tests, or docs over summaries."
+  ],
+  "whenToAvoid": [
+    "You only need to explore or plan.",
+    "There is no concrete evidence to collect yet.",
+    "The task is implementation rather than proof gathering."
+  ],
+  "caseId": "evidence-pack",
+  "source": "bootstrap-from-skill-md"
+}

package/skills/shared/framework-guide/SKILL.md

@@ -0,0 +1,28 @@
+---
+name: framework-guide
+description: Load and apply repository framework conventions from rules and overlays.
+---
+
+# Framework Guide
+
+## Purpose
+
+Help the agent find the right rules and overlays for the current stack or project.
+
+## When to Use
+
+- The repository uses framework-specific conventions.
+- You need to understand which overlay or rule set applies.
+- The stack rules are more specific than the shared core rules.
+
+## How
+
+- Identify the relevant stack, load the matching overlay, and note the non-negotiable conventions.
+- Explain the rules at the level needed to keep work aligned without rewriting the whole rule tree.
+- Call out any stack-specific boundary, naming, or response-shape conventions that other skills must honor.
+
+## When to Avoid
+
+- The change is already fully constrained by the active rules.
+- You need business logic or implementation work instead of framework guidance.
+- The task only needs repo exploration or evidence collection.

package/skills/shared/framework-guide/assets/golden-prompt-suite.current.js

@@ -0,0 +1,22 @@
+#!/usr/bin/env node
+
+module.exports = {
+  "name": "framework-guide.golden-prompt-suite.current",
+  "skill": "framework-guide",
+  "description": "Load and apply repository framework conventions from rules and overlays.",
+  "cases": [
+    {
+      "id": "framework-governance-overview",
+      "prompt": "Explain how the repository conventions, rules, and skill layers fit together.",
+      "expectedFirstSkill": "framework-guide",
+      "expectedOutput": [
+        "guidance",
+        "rules",
+        "usage notes"
+      ],
+      "expectedStop": "Do not invent repo policy.",
+      "expectedHandoff": null
+    }
+  ]
+};
+

package/skills/shared/framework-guide/references/README.md

@@ -0,0 +1,17 @@
+# framework-guide References
+
+This directory is local to the skill and is the only reference entry point for this skill.
+
+## Contents
+
+- `../skill-spec.json`: design source for regenerating this skill
+- `prompt-template.md`: reusable prompt wording for this skill
+- `cases.md`: human-readable replay case for this skill
+- `../assets/golden-prompt-suite.current.js`: machine-readable local case fixture
+
+## Guidance
+
+- Keep skill-specific explanation here instead of pushing it into a stage-level shared library.
+- Keep shared hard constraints in `../../../rules/index.md`.
+- Keep canonical names unchanged.
+- Skill summary: Load and apply repository framework conventions from rules and overlays.

package/skills/shared/framework-guide/references/cases.md

@@ -0,0 +1,26 @@
+# framework-guide
+
+## Replay Case
+
+- Case: `framework-governance-overview`
+- First skill: `framework-guide`
+
+### Prompt
+
+```text
+Explain how the repository conventions, rules, and skill layers fit together.
+```
+
+### Expected Output
+
+- guidance
+- rules
+- usage notes
+
+### Stop Rule
+
+- Do not invent repo policy.
+
+### Handoff
+
+- none

package/skills/shared/framework-guide/references/prompt-template.md

@@ -0,0 +1,18 @@
+# framework-guide
+
+## Prompt Template
+
+```text
+You are handling the framework-guide skill.
+Goal: Load and apply repository framework conventions from rules and overlays
+Inputs: the request, supporting evidence, constraints, and existing repository facts.
+Output contract: guidance, rules, usage notes.
+Stop rule: Do not invent repo policy.
+Handoff: none.
+Keep the response scoped to this stage and do not absorb the work of the next stage.
+```
+
+## Notes
+
+- Keep the template aligned with the local replay case and local assets.
+- Read `cases.md` for the matching replay example.

package/skills/shared/framework-guide/skill-spec.json

@@ -0,0 +1,24 @@
+{
+  "group": "shared",
+  "name": "framework-guide",
+  "description": "Load and apply repository framework conventions from rules and overlays.",
+  "title": "Framework Guide",
+  "purpose": "Help the agent find the right rules and overlays for the current stack or project.",
+  "whenToUse": [
+    "The repository uses framework-specific conventions.",
+    "You need to understand which overlay or rule set applies.",
+    "The stack rules are more specific than the shared core rules."
+  ],
+  "how": [
+    "Identify the relevant stack, load the matching overlay, and note the non-negotiable conventions.",
+    "Explain the rules at the level needed to keep work aligned without rewriting the whole rule tree.",
+    "Call out any stack-specific boundary, naming, or response-shape conventions that other skills must honor."
+  ],
+  "whenToAvoid": [
+    "The change is already fully constrained by the active rules.",
+    "You need business logic or implementation work instead of framework guidance.",
+    "The task only needs repo exploration or evidence collection."
+  ],
+  "caseId": "framework-governance-overview",
+  "source": "bootstrap-from-skill-md"
+}