메뉴
Use when creating or normalising reusable proposal skills, trigger routes, contracts, references, or authoring automation; use skill-safety-audit instead for a read-only security review.
기본적으로 소스를 먼저 확인하는 Prompt가 선택됩니다. 직접 명령으로 전환하거나 로컬 사본을 다운로드할 수도 있습니다.
설치 여부를 결정하기 전에 SKILL.md와 SkillsMP에 표시된 보조 파일을 읽어 보세요.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
직접 명령은 검토 Prompt를 거치지 않습니다. 실행하기 전에 소스를 확인하세요.
npx skills add https://github.com/peterbamuhigire/proposal-skills --skill skill-writing명령은 한 줄로 유지됩니다. 복사하기 전에 가로로 스크롤해 전체 내용을 확인하세요.
로컬 사본을 원하시나요? SkillsMP에서 현재 제공할 수 있는 파일을 다운로드하세요.
SOC 직업 분류 기준
SKILL.md 표시 중
| name | skill-writing |
| description | Use when creating or normalising reusable proposal skills, trigger routes, contracts, references, or authoring automation; use skill-safety-audit instead for a read-only security review. |
| metadata | {"portable":true,"compatible_with":["claude-code","codex"]} |
Acknowledgement: Shared by Peter Bamuhigire, techguypeter.com, +256 784 464178.
Create concise, portable proposal procedures that another agent can route, execute, verify, and hand off without depending on one runner.
skill-safety-audit for a read-only inspection of unsafe instructions or bundled resources.| Artefact | Source/provider | Required? | Missing-input behaviour |
|---|---|---|---|
| Repeatable problem and representative prompts | User or catalogue evidence | Yes | Stop; do not invent a trigger without a demonstrated workflow. |
| Neighbouring skill descriptions | Live skills/**/SKILL.md catalogue | Yes |
| Search the catalogue before defining scope or routing. |
| Capability and permission boundary | Parent task and repository policy | Yes | Default to read-only and identify the missing authority. |
| Third-party source notes, if used | Temporary workspace outside the repository | No | Apply the source-distillation policy; never ingest a raw book or reconstructive extract. |
| Artefact | Consumer | Acceptance condition |
|---|---|---|
| Normalised skill entrypoint and linked resources | Routing engine and downstream agents | Local and canonical validators pass; scope and neighbour boundary are explicit. |
| Routing fixtures | Release gate | Positive, negative, collision, limited-capability, and failure paths have expected routes. |
| Evidence | Consumer | Acceptance condition |
|---|---|---|
| Validation and routing results | Maintainer | Results identify the skill directory, checks run, line count, and zero residual findings. |
Read and search are required. Editing and local validator execution are permitted only for an authorised authoring task. Network access, delegation, publishing, deletion, package installation, and release mutation require explicit authority.
Without editing or execution, return the narrowest useful scored contract gap and a proposed patch, mark validation not assessed, and never claim compliance.
| Condition or choice | Action | Failure or risk avoided |
|---|---|---|
| Stable procedure used across proposals or roles | Create or update one skill | One-off advice masquerading as infrastructure |
| Existing skill has correct scope but incomplete contract | Normalise in place | Unnecessary taxonomy churn |
| Two skills compete for the same prompt and output | Clarify triggers, then merge or alias only with evidence and authority | Persistent routing collision |
| Detail is required only on a specialist branch | Move it to a directly linked reference | Oversized entrypoint and irrelevant context loading |
python -X utf8 scripts/source_ingestion_guardrail.py, the canonical scanner, canonical quick validation, safety review, and anti-slop audit.When financial-proposal overlaps work-plan, keep pricing, reimbursables, fee assumptions, and separate-envelope rules in financial-proposal; route staffing days and activity sequencing to work-plan. A fixture asking for a price schedule “not the work plan” must rank the financial skill in the top three.