agentleFS
MicrosoftKnown publishermicrosoft/powerplatform-build-tools/.claude/skills/workitem/SKILL.md

workitem

Skill314 starsChanged 7 months agoMITView on GitHub

Read an ADO or GitHub work item, research similar past PRs, implement the fix, and leave the branch ready for /create-pr.

The file

Show its rules and text

We split this skill into the rules it gives an agent

21 rules. About 0 are common, 11 are distinctive, and 10 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈11

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 10

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Use this skill to read an ADO or GitHub work item, understand its impact on this codebase, research similar past PRs and work items, then …

    line 10
  2. Distinctive:

    Before researching or implementing, ensure the knowledge base is current.

    line 21
  3. Distinctive:

    If last sync was ≤ 7 days ago: skip, proceed to Step 1

    line 27
  4. Very specific:

    If last sync was > 7 days ago or file missing: run /knowledge-sync inline now, then continue

    line 28
  5. Distinctive:

    If only a URL is given, extract the ID and detect the type from the URL pattern:

    line 50
  6. Distinctive:

    Look for: how the fix was structured, which files were changed, patterns used.

    line 73
  7. Very specific:

    Read memory/ado-knowledge.md for any entry matching the work item topic.

    line 93
  8. Very specific:

    Check MEMORY.md known customer issues and hard rules.

    line 94
  9. Distinctive:

    Before writing any code, print:

    line 98
  10. Distinctive:

    If a confirmed fix already exists for the exact same issue in a merged PR, reuse that pattern directly rather than solving from scratch.

    line 108
  11. Distinctive:

    Before touching code, map the work item to the correct layer using the architecture below.

    line 115
  12. Distinctive:

    Read the relevant source files to understand the current behaviour.

    line 116
  13. Very specific:

    All 32 tasks share the same host abstraction (BuildToolsHost, BuildToolsRunnerParams) — changes there affect every task

    line 141
  14. Distinctive:

    Always fetch the latest main before branching:

    line 150
  15. Very specific:

    New task input: add to task.json inputs array AND handle in index.ts via parameterMap

    line 177
  16. Very specific:

    Auth issue: check getCredentials.ts and the cloudInstance resolution map

    line 178
  17. Very specific:

    Environment URL not found: check the 4-level fallback chain in getEnvironmentUrl.ts

    line 179
  18. Very specific:

    Build/pack issue: check gulp/pack.mjs and webpack.config.js

    line 181
  19. Very specific:

    Dependency vulnerability: use /fix-dependencies

    line 182
  20. Distinctive:

    After implementing, always run:

    line 184
  21. Distinctive:

    Leave the branch in a clean, pushed state and note:

    line 206

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Work Item — Read, Triage, and Implement

Use this skill to read an ADO or GitHub work item, understand its impact on this codebase, research similar past PRs and work items, then implement and leave the branch ready for /create-pr.

Invoke as: /workitem <url-or-id>

Zero user input required. Research first, implement second — never skip the research phase.


Step 0 — Sync knowledge if stale

Before researching or implementing, ensure the knowledge base is current.

grep "## Last sync" memory/ado-knowledge.md 2>/dev/null | tail -1
  • If last sync was ≤ 7 days ago: skip, proceed to Step 1
  • If last sync was > 7 days ago or file missing: run /knowledge-sync inline now, then continue

This ensures known patterns, confirmed fixes, and accepted risks are up to date before you base any implementation on them.


Step 1 — Fetch the work item

GitHub issue:

gh issue view <number> 2>&1
gh issue view <number> --comments 2>&1

ADO work item (requires az CLI):

az boards work-item show --id <id> 2>&1

If only a URL is given, extract the ID and detect the type from the URL pattern:

  • github.com/.../issues/<n> → GitHub issue
  • dev.azure.com/.../workitems/<n> or /_workitems/edit/<n> → ADO work item

Step 2 — Research similar past work (always before touching code)

Extract 2-4 keywords from the work item title/description, then run all three in parallel:

2a — Search merged PRs for similar fixes

gh search prs --repo microsoft/powerplatform-build-tools \
  --state merged \
  --limit 10 \
  "<keyword1> <keyword2>" 2>&1

# For the top 3 most relevant results, read the full body:
gh pr view <past-pr-number> --json title,body,mergedAt,files 2>&1

Look for: how the fix was structured, which files were changed, patterns used.

2b — Search ADO for similar resolved items

az boards query --wiql "
  SELECT [System.Id], [System.Title], [System.State], [System.ChangedDate]
  FROM WorkItems
  WHERE [System.AreaPath] UNDER 'OneCRM\Client\UnifiedClient\AppLifeCycle\PPBT Extensions'
    AND [System.State] IN ('Closed', 'Resolved', 'Done')
    AND [System.Title] CONTAINS '<keyword>'
  ORDER BY [System.ChangedDate] DESC
" --output json 2>&1

# Fetch details for most relevant matches:
az boards work-item show --id <id> 2>&1

2c — Check local knowledge base

Read memory/ado-knowledge.md for any entry matching the work item topic. Check MEMORY.md known customer issues and hard rules.

Research summary (print before proceeding)

Before writing any code, print:

Research findings:
- Similar PRs: <PR titles + numbers, or "none">
- Similar ADO items: <IDs + titles, or "none">
- Applicable hard rules: <from MEMORY.md, or "none">
- Reusable pattern: <describe how the fix should mirror a past fix, or "novel">

If a confirmed fix already exists for the exact same issue in a merged PR, reuse that pattern directly rather than solving from scratch.


Step 3 — Understand impact using architecture (read-only, before branching)

Before touching code, map the work item to the correct layer using the architecture below. Read the relevant source files to understand the current behaviour.

Layer map — where to look:

Work item topicLayerKey files to read
Task input/output, pipeline variable, service connectionbuild-toolssrc/tasks/<name>/index.ts, src/params/auth/getCredentials.ts, src/params/auth/getEnvironmentUrl.ts
pac CLI argument, action logic, auth flowcli-wrappernode_modules/@microsoft/powerplatform-cli-wrapper/dist/actions/<action>.ts
New/changed task definitionbuild-toolssrc/tasks/<name>/task.json, src/tasks/<name>/index.ts
Build, packaging, VSIX, stage GUIDsbuild systemgulp/pack.mjs, extension/task-metadata.json, extension/extension-manifest.json
Vulnerability / dependencydependenciespackage.json overrides, package-lock.json — use /fix-dependencies
Test failureteststest/unit-test/, test/functional-test/

Architecture call chain (for context):

Azure DevOps Pipeline
  → src/tasks/<name>/index.ts          (build-tools: reads inputs, calls cli-wrapper)
  → cli-wrapper: actions/<action>.ts   (assembles pac args, handles auth)
  → pac CLI binary (pac.exe)           (executes, spawned via child_process)
  → Power Platform API

Key constraints:

  • All 32 tasks share the same host abstraction (BuildToolsHost, BuildToolsRunnerParams) — changes there affect every task
  • task.json declares the Azure DevOps UI contract — changing input names is breaking for existing pipelines
  • bundleDependencies packages (azure-pipelines-task-lib, @microsoft/powerplatform-cli-wrapper, fs-extra, semver) ship inside the VSIX — their interfaces are the extension's API surface
  • Four release stages (LIVE / BETA / DEV / EXPERIMENTAL) each have unique task GUIDs in task-metadata.json

Step 4 — Sync with origin and create a branch

Always fetch the latest main before branching:

git fetch origin 2>&1
git checkout main 2>&1
git merge origin/main --no-edit 2>&1

Then create the branch following the project naming convention:

git checkout -b users/<your-alias>/<short-description>
# e.g. users/jbujula/fix-export-solution-timeout

Reference the work item ID in the branch name if it's a tracked ticket:

git checkout -b users/<alias>/<id>-<short-description>

Step 5 — Implement

Make the minimal change that resolves the work item. Follow existing patterns:

  • New task input: add to task.json inputs array AND handle in index.ts via parameterMap
  • Auth issue: check getCredentials.ts and the cloudInstance resolution map
  • Environment URL not found: check the 4-level fallback chain in getEnvironmentUrl.ts
  • pac CLI arg change: this is in cli-wrapper, not this repo — raise issue upstream; workaround via BuildToolsRunnerParams if urgent
  • Build/pack issue: check gulp/pack.mjs and webpack.config.js
  • Dependency vulnerability: use /fix-dependencies

After implementing, always run:

npm run build 2>&1   # must pass before committing
npm test 2>&1        # must pass before committing

Step 6 — Commit

Stage relevant files only (never git add .):

git add src/tasks/<name>/... package.json package-lock.json  # etc.
git commit -m "fix: <short description> (#<work-item-id>)"

Step 7 — Hand off to /create-pr

Leave the branch in a clean, pushed state and note:

  • Work item URL
  • What was changed and why
  • Any follow-up items or known limitations

Then run /create-pr to create the pull request.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.