Define requirements for an OpenClaw integration
Document a proposed integration's data needs, actions, credentials, and review boundaries.
Use when the user wants to create or update a Team Topologies Team API document with AI literacy portfolio assessment data — generates a template Team API with literacy levels, discipline scores, shared gaps, and improvement plans, or updates an existing Team API with the latest assessment data
--- name: team-api-d13180db2ef9 description: Use when the user wants to create or update a Team Topologies Team API document with AI literacy portfolio assessment data — generates a template Team API with literacy levels, discipline scores, shared gaps, and improvement plans, or updates an existing Team API with the latest assessment data --- # Team API Generate or update a Team Topologies Team API document enriched with AI literacy portfolio assessment data. The Team API is a document that describes a team's capabilities, communication preferences, and service offerings — adding AI literacy data makes the team's engineering maturity visible to the rest of the organisation. This skill works in two modes: - **Update mode**: reads an existing Team API document and adds or refreshes the AI literacy section using the latest portfolio assessment data - **Template mode**: generates a new Team API document from the template in `references/team-api-template.md`, populated with portfolio assessment data ## Input The skill needs: - **Portfolio assessment file** — the most recent `assessments/*-portfolio-assessment.md`. If none exists, tell the user to run `/portfolio-assess` first. - **Existing Team API file** (optional) — if the user provides a path to an existing Team API document, update it. If not, generate from the template. - **Team directory** (optional) — where to write the output. Defaults to the current directory. ## Process ### Step 1: Find Portfolio Assessment Data ```bash ls assessments/*-portfolio-assessment.md 2>/dev/null | sort | tail -1 ``` If no file exists, stop and suggest running `/portfolio-assess`. Read the most recent portfolio assessment and extract: - Portfolio median level - Assessment coverage percentage - Repo detail table (each repo with level and discipline scores) - Shared gaps - Improvement plan (organisation-wide items) - Next assessment date ### Step 2: Check for Existing Team API Ask the user: ```text Do you have an existing Team API document to update? 1. Yes — provide the file path 2. No — generate a new one from the template ``` ### Step 3a: Update Existing Team API If the user provides an existing file, read it and look for an `## AI Literacy` or `## AI Engineering Maturity` section. If found, replace it with fresh data. If not found, append the section before the last section of the document (typically "Further Information" or similar). The AI Literacy section to insert: ```markdown ## AI Literacy **Portfolio median level**: LN — Level Name **Assessment coverage**: N% of repos fully assessed **Last portfolio assessment**: YYYY-MM-DD **Next assessment due**: YYYY-MM-DD ### Repo Levels | Repo | Level | Confidence | Last Assessed | | --- | --- | --- | --- | | repo-name | LN | assessed/estimated | YYYY-MM-DD | ### Discipline Scores (assessed repos only) | Repo | Context | Constraints | Guardrails | | --- | --- | --- | --- | | repo-name | N/5 | N/5 | N/5 | ### Active Shared Gaps | Gap | Repos Affected | Recommended Action | | --- | --- | --- | | gap description | N/M | action | ### Current Improvement Focus [Organisation-wide improvement items from the portfolio assessment, presented as the team's current AI engineering priorities] ``` Preserve all other sections of the existing Team API untouched. ### Step 3b: Generate New Team API from Template If no existing file, read `references/team-api-template.md` and populate it with: - Team name (ask the user) - Team type (ask: stream-aligned, enabling, complicated-subsystem, or platform) - The AI Literacy section populated from portfolio assessment data - Placeholder sections for the user to fill in (communication preferences, service offerings, dependencies) Write to the team directory as `team-api.md`. ### Step 4: Summary Report what was done: ```text Team API [created/updated]: path/to/team-api.md AI Literacy section: populated from portfolio assessment (YYYY-MM-DD) Portfolio median: LN Repos covered: N Shared gaps: N Next assessment: YYYY-MM-DD ``` ## What Gets Pulled from Portfolio Assessment | Portfolio field | Team API field | | --- | --- | | Portfolio median level | AI Literacy header | | Assessment coverage | AI Literacy header | | Repo detail table | Repo Levels table (simplified) | | Discipline scores | Discipline Scores table | | Shared gaps | Active Shared Gaps table | | Organisation-wide improvements | Current Improvement Focus | | Next assessment date | Next assessment due | ## What This Skill Does NOT Do - Does not run `/portfolio-assess` — the data must already exist - Does not modify the portfolio assessment — it reads, not writes - Does not fill in non-AI sections of the Team API — those are the team's responsibility - Does not enforce Team Topologies practices — it provides the document structure, the team decides how to use it--- name: team-api-d13180db2ef9 description: Use when the user wants to create or update a Team Topologies Team API document with AI literacy portfolio assessment data — generates a template Team API with literacy levels, discipline scores, shared gaps, and improvement plans, or updates an existing Team API with the latest assessment data --- # Team API Generate or update a Team Topologies Team API document enriched with AI literacy portfolio assessment data. The Team API is a document that describes a team's capabilities, communication preferences, and service offerings — adding AI literacy data makes the team's engineering maturity visible to the rest of the organisation. This skill works in two modes: - **Update mode**: reads an existing Team API document and adds or refreshes the AI literacy section using the latest portfolio assessment data - **Template mode**: generates a new Team API document from the template in `references/team-api-template.md`, populated with portfolio assessment data ## Input The skill needs: - **Portfolio assessment file** — the most recent `assessments/*-portfolio-assessment.md`. If none exists, tell the user to run `/portfolio-assess` first. - **Existing Team API file** (optional) — if the user provides a path to an existing Team API document, update it. If not, generate from the template. - **Team directory** (optional) — where to write the output. Defaults to the current directory. ## Process ### Step 1: Find Portfolio Assessment Data ```bash ls assessments/*-portfolio-assessment.md 2>/dev/null | sort | tail -1 ``` If no file exists, stop and suggest running `/portfolio-assess`. Read the most recent portfolio assessment and extract: - Portfolio median level - Assessment coverage percentage - Repo detail table (each repo with level and discipline scores) - Shared gaps - Improvement plan (organisation-wide items) - Next assessment date ### Step 2: Check for Existing Team API Ask the user: ```text Do you have an existing Team API document to update? 1. Yes — provide the file path 2. No — generate a new one from the template ``` ### Step 3a: Update Existing Team API If the user provides an existing file, read it and look for an `## AI Literacy` or `## AI Engineering Maturity` section. If found, replace it with fresh data. If not found, append the section before the last section of the document (typically "Further Information" or similar). The AI Literacy section to insert: ```markdown ## AI Literacy **Portfolio median level**: LN — Level Name **Assessment coverage**: N% of repos fully assessed **Last portfolio assessment**: YYYY-MM-DD **Next assessment due**: YYYY-MM-DD ### Repo Levels | Repo | Level | Confidence | Last Assessed | | --- | --- | --- | --- | | repo-name | LN | assessed/estimated | YYYY-MM-DD | ### Discipline Scores (assessed repos only) | Repo | Context | Constraints | Guardrails | | --- | --- | --- | --- | | repo-name | N/5 | N/5 | N/5 | ### Active Shared Gaps | Gap | Repos Affected | Recommended Action | | --- | --- | --- | | gap description | N/M | action | ### Current Improvement Focus [Organisation-wide improvement items from the portfolio assessment, presented as the team's current AI engineering priorities] ``` Preserve all other sections of the existing Team API untouched. ### Step 3b: Generate New Team API from Template If no existing file, read `references/team-api-template.md` and populate it with: - Team name (ask the user) - Team type (ask: stream-aligned, enabling, complicated-subsystem, or platform) - The AI Literacy section populated from portfolio assessment data - Placeholder sections for the user to fill in (communication preferences, service offerings, dependencies) Write to the team directory as `team-api.md`. ### Step 4: Summary Report what was done: ```text Team API [created/updated]: path/to/team-api.md AI Literacy section: populated from portfolio assessment (YYYY-MM-DD) Portfolio median: LN Repos covered: N Shared gaps: N Next assessment: YYYY-MM-DD ``` ## What Gets Pulled from Portfolio Assessment | Portfolio field | Team API field | | --- | --- | | Portfolio median level | AI Literacy header | | Assessment coverage | AI Literacy header | | Repo detail table | Repo Levels table (simplified) | | Discipline scores | Discipline Scores table | | Shared gaps | Active Shared Gaps table | | Organisation-wide improvements | Current Improvement Focus | | Next assessment date | Next assessment due | ## What This Skill Does NOT Do - Does not run `/portfolio-assess` — the data must already exist - Does not modify the portfolio assessment — it reads, not writes - Does not fill in non-AI sections of the Team API — those are the team's responsibility - Does not enforce Team Topologies practices — it provides the document structure, the team decides how to use it
Imported from the listed AgentSkill.sh category. Complete supplied package files and notices are retained. Required external tools and credentials are not installed. No bundled code was executed during import. Individual agent task behavior has not been tested.