一键导入
kramme-siw-init
Initialize structured implementation workflow documents in siw/ (spec, LOG.md, issues)
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
Initialize structured implementation workflow documents in siw/ (spec, LOG.md, issues)
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
| name | kramme:siw:init |
| description | Initialize structured implementation workflow documents in siw/ (spec, LOG.md, issues) |
| argument-hint | [spec-file(s) | folder | discover] [--auto] |
| disable-model-invocation | true |
| user-invocable | true |
Set up the three-document system for tracking complex implementations locally, without requiring Linear or other external issue trackers.
This command ONLY initializes workflow documents.
Issue definition is a separate workflow. After this command completes, invoke /kramme:siw:issue-define to create your first issue, or /kramme:siw:generate-phases when the spec should be decomposed into phased issues. Spec hardening is a separate workflow. To strengthen an existing SIW spec, use /kramme:siw:discovery.
Use this shared vocabulary in setup summaries and handoffs:
product-only: the source explains the problem, user, or desired outcome but still lacks testable scope or success criteria.requirements-only: scope and success criteria exist, but implementation planning still lacks phase boundaries, dependencies, or enough technical context.planning-ready: the SIW spec is concrete enough for /kramme:siw:generate-phases or /kramme:siw:issue-define.implementation-ready: an issue file has executable scope, dependencies, acceptance criteria, and verification. This skill never produces implementation-ready artifacts.siw:init creates the tracked planning container. It may preserve a product-only or requirements-only source as context, or produce a planning-ready SIW spec when the imported/discovered content is concrete enough; it must not imply implementation is ready until issues exist.
/kramme:siw:init [spec-file(s) | folder | discover]
↓
[Check for existing files] -> Found? -> Ask: resume or start fresh
↓
[Handle arguments] -> file/folder: import content
-> discover: run greenfield discovery, then import siw/DISCOVERY_BRIEF.md
-> empty: continue to brief interview
↓
[Brief interview OR Confirm imported content]
↓
[Read STRATEGY.md if present]
↓
[Select work context] -> Profile for downstream tool adaptation
↓
[Auto-detect spec type] -> Confirm filename
↓
[Ask about supporting specs] -> Need detailed specs?
↓
[Create documents] -> siw/spec, siw/LOG.md, siw/issues/, (supporting-specs/)
↓
[Report success] -> Suggest next readiness-gated skill
Before Phase 1, parse $ARGUMENTS as shell-style arguments. If --auto is present, set AUTO_MODE=true and remove it from the remaining input. --auto uses safe initialization defaults when enough context is already supplied: keep linked files in place, use all discovered files from an explicit folder, choose the auto-detected work context and spec filename, and create a single spec file without supporting specs. It does not bypass required project context, existing-workflow protection, deletion confirmation, file overwrite checks, or fresh discovery topic requirements.
Check if any workflow files already exist:
find siw -maxdepth 1 \( \
-name "LOG.md" -o \
-name "OPEN_ISSUES_OVERVIEW.md" -o \
-name "AUDIT_IMPLEMENTATION_REPORT.md" -o \
-name "AUDIT_SPEC_REPORT.md" -o \
-name "PRODUCT_AUDIT.md" -o \
-name "SIW_*.md" -o \
-name "SPEC_STRENGTHENING_PLAN.md" -o \
-name "DISCOVERY_BRIEF.md" -o \
-name "issues" \
\) -print 2> /dev/null
# Permanent SIW spec detection (referenced as `permanent-spec find` elsewhere in this skill).
# Case-insensitive so lowercase/mixed-case filenames like `feature_spec.md` are not missed.
# Synced SIW spec-exclusion contract (keep aligned across SIW spec detectors): `LOG.md`, `OPEN_ISSUES_OVERVIEW.md`, `DISCOVERY_BRIEF.md`, `SPEC_STRENGTHENING_PLAN.md`, `AUDIT_*.md`, `PRODUCT_AUDIT.md`, `SIW_*.md`.
find siw -maxdepth 1 -type f \( -iname "*SPEC*.md" -o -iname "*SPECIFICATION*.md" -o -iname "*PLAN*.md" -o -iname "*DESIGN*.md" \) \
! -iname "LOG.md" \
! -iname "OPEN_ISSUES_OVERVIEW.md" \
! -iname "DISCOVERY_BRIEF.md" \
! -iname "SPEC_STRENGTHENING_PLAN.md" \
! -iname "AUDIT_*.md" \
! -iname "PRODUCT_AUDIT.md" \
! -iname "SIW_*.md" \
2> /dev/null
Read references/existing-workflow-handling.md and follow the first matching branch for siw/DISCOVERY_BRIEF.md only, siw/DISCOVERY_BRIEF.md + siw/SPEC_STRENGTHENING_PLAN.md, siw/SPEC_STRENGTHENING_PLAN.md only, other workflow files, or no files.
Read references/argument-handling.md and follow the matching case for file paths, folder paths, discover / interview, or no arguments. This phase sets linked_spec_files, linked_spec_readiness_context, discovered_content, project_description, or resolved_arguments for later phases.
Skip this phase if imported_spec_content or discovered_content exists from Phase 1.5.
If AUTO_MODE=true and Phase 2 would run, stop before asking with MISSING REQUIREMENT: project context required for --auto; pass a spec file, folder, discovery brief, or use discover <topic>.
Read the Phase 2 templates from references/interviews.md, use them to gather project context, urgency and outcome, scope boundaries, and decision boundaries, then store the responses as project_description, why_now, out_of_scope_non_goals, and decision_boundaries_notes.
Only executed if linked_spec_files exists from Phase 1.5 (file/folder import).
Skip this phase if discovered_content exists (discover mode already has confirmation built in).
Show the files that will be linked:
Linked Specification Files:
───────────────────────────
The following files will be referenced (not duplicated) in the SIW spec:
1. {file1} - "{title from first heading}"
2. {file2} - "{title from first heading}"
...
These files remain the source of truth. The SIW spec will link to them.
If AUTO_MODE=true, choose Keep in place. Otherwise read the Phase 2.5 file-location template from references/interviews.md and use AskUserQuestion.
For both "Move to siw/" and "Copy to siw/", before transferring each file, check the target path with [ -e "siw/{filename}" ]. If a file already exists at the target, read the Phase 2.5 collision template from references/interviews.md and use AskUserQuestion before overwriting.
Apply the user's choice per file, then:
linked_spec_files, continue to Phase 2.6siw/{filename} (or the renamed/skipped path per the collision prompt above)linked_spec_files with the resulting pathssiw/ (or the renamed/skipped path per the collision prompt above)linked_spec_files with the resulting pathsOnly executed if linked_spec_files exists.
If AUTO_MODE=true, use the inferred default value. Otherwise read the Phase 2.6 project-context template from references/interviews.md and use AskUserQuestion.
Store the response as project_description.
If user provides empty response or selects "Skip", use a generic description derived from the linked file names.
Before selecting the work context, check for repo-root STRATEGY.md. If it exists, read it, store concise strategy_context, mark stale context with STALE: when last_updated is older than 90 days, and surface conflicts with active tracks or non-goals without blocking initialization. If no strategy exists, stay silent for narrow work; for broad product-direction work, include MISSING PRODUCT CONTEXT: no STRATEGY.md found in the generated spec's strategy context note and suggest /kramme:product:strategy after initialization.
Select a work context profile that tells downstream tools (spec-audit, product-review, discovery, generate-phases) how to adapt their rigor and focus.
Read the profile definitions and auto-detection heuristics from references/work-context-profiles.md.
Based on project_description or discovered_content topic, use the keyword heuristics from the reference file to suggest a profile. If AUTO_MODE=true, choose the auto-detected profile; otherwise read the Phase 2.8 work-context template from references/interviews.md, deduplicate options as documented there, and use AskUserQuestion.
Store the selected profile as work_context_profile with all attribute values from the reference file (work_type, maturity, priority_dimensions, deprioritized, notes).
Use the Phase 3 filename detection heuristics in references/interviews.md to auto-detect the most appropriate spec filename. If AUTO_MODE=true, choose it automatically. Otherwise read the Phase 3 specification-document template from references/interviews.md and use AskUserQuestion; if "Custom name" is selected, use AskUserQuestion to get the filename.
Store as spec_filename.
If AUTO_MODE=true, choose No - single spec file is enough. Otherwise read the Phase 3.5 supporting-specifications template from references/interviews.md and use AskUserQuestion.
Store as use_supporting_specs.
Read references/document-creation.md, apply its readiness classification procedure, and follow it to create siw/, siw/{spec_filename}, siw/LOG.md, siw/OPEN_ISSUES_OVERVIEW.md, siw/issues/, and optional siw/supporting-specs/.
When strategy_context exists, include a concise ## Product Strategy Context section in the generated SIW spec using the relevant placeholders from assets/spec-templates.md. Keep it to strategy facts and conflicts; do not duplicate the entire STRATEGY.md.
Read the Phase 5 success-report templates from references/success-report.md, display the applicable summary sections, then stop.
Include one readiness line in the success report:
Artifact readiness: <product-only|requirements-only|planning-ready> — <one-line reason and next skill>
If the result is product-only or requirements-only, point to /kramme:siw:discovery before phase or issue creation. If it is planning-ready, point to /kramme:siw:generate-phases for phased work or /kramme:siw:issue-define for a single first issue.
STOP HERE. Wait for the user's next instruction.
Analyze code quality of branch changes using specialized review agents (tests, errors, types, security, performance, slop, lean deletion, refactor fit, simplification). Outputs REVIEW_OVERVIEW.md with actionable findings, or replies inline with --inline. Use --team for multi-agent cross-validation. Not for UX, visual, or accessibility review -- use kramme:pr:ux-review for those.
Resolve findings from code reviews by implementing fixes and documenting changes. Implements fixes as commits on the current branch. Use --team to resolve independent findings in parallel by file area.
Charts huge or foggy initiatives into a local `.context` decision map and resolves one typed frontier ticket per session until the work is ready for SIW or another execution workflow. Use when the route to a destination cannot fit in one agent session or parallel workspaces need coordinated planning state. Not for clear specs, ordinary issue decomposition, implementation, or Linear-native tracking.
Builds a clearly throwaway logic/state or UI prototype to answer one design question before implementation hardens. Use when the user wants to sanity-check a state model, data shape, API surface, page layout, component direction, or interaction idea with disposable code. Not for production implementation, polished demos, visual diff reports, permanent routes, or broad design-system work.
Investigates a question against primary sources and saves one cited Markdown artifact. Use for reading legwork: official docs/API facts, source-code or spec checks, standards, and first-party service behavior before planning or implementation. Not for making product or architecture decisions, implementing code, broad web search, secondary blog summaries, or uncited answers.
Break spec into atomic, phase-based issues with tests and validation