adk-implement
Implement, build, add, write, ship, code-up, wire-up code in an existing repo. Triggers on: Jira URL or KEY-NUM (specialized from-jira sub-flow — the most common path), GitHub issue URL or #N (from-issue), Confluence TDD URL (from-tdd), Slack thread permalink (from-slack-thread), or a freeform description (greenfield). Git mandatory; GitHub MCP optional but enables PR-by-URL flow. Every run: question-first (3 questions max), advisor-strategy plan with 2–4 trade-off options, edit-format discipline (SEARCH/REPLACE blocks per shared/edit-format.md), repo-native typecheck+lint+narrow-tests on each checkpoint. Writes plan/steps/diffs/report under <repo>/.temp/adk/implement/<task>/. Pushes only after explicit confirmation; never force-pushes; never merges; never touches a protected branch. Pulls Jira context via adk-mcp-atlassian, GitHub context via adk-mcp-github, and optional RAG via adk-mcp-rag. Supports —plan (read-only planning) → —act (writes). Sub-flows under references/.
Source
skills/adk-implement/SKILL.md
Frontmatter
name: adk-implementdescription: | Implement, build, add, write, ship, code-up, wire-up code in an existing repo. Triggers on: Jira URL or KEY-NUM (specialized from-jira sub-flow — the most common path), GitHub issue URL or #N (from-issue), Confluence TDD URL (from-tdd), Slack thread permalink (from-slack-thread), or a freeform description (greenfield). Git mandatory; GitHub MCP optional but enables PR-by-URL flow. Every run: question-first (3 questions max), advisor-strategy plan with 2–4 trade-off options, edit-format discipline (SEARCH/REPLACE blocks per shared/edit-format.md), repo-native typecheck+lint+narrow-tests on each checkpoint. Writes plan/steps/diffs/report under `<repo>/.temp/adk/implement/<task>/`. Pushes only after explicit confirmation; never force-pushes; never merges; never touches a protected branch. Pulls Jira context via adk-mcp-atlassian, GitHub context via adk-mcp-github, and optional RAG via adk-mcp-rag. Supports --plan (read-only planning) → --act (writes). Sub-flows under references/.allowed-tools: [Read, Edit, Write, Grep, Glob, Bash, WebFetch, Agent]argument-hint: "<input-url-or-description> [-i|--interactive] [--plan|--act] [--scope vertical-slice|full|spike] [--no-push] [--detailed] [--deep]"metadata: category: code kind: task layer: 1 paths: ["**/*.{ts,tsx,js,jsx,py,go,rs,java,rb,php,cs,kt,swift,c,cpp,h,hpp,sh,sql,yaml,yml,json,toml}"] model: sonnet effort: medium user-invocable: true disable-model-invocation: false needs_mcp_required: [adk-mcp-github] needs_mcp_optional: [adk-mcp-atlassian, adk-mcp-slack, adk-mcp-rag] needs_meta_info: [workspaces, repos] forks_emitted: [scope, approach, test-framework, pr-strategy, commit-style, linter-tolerance, breaking-change-policy, mode, model-depth]Workflow body
adk-implement — write code from any input
Polymorphic on input. Git mandatory; GitHub MCP optional (enables PR-by-URL flow).
Repo-bound skill — must run from inside a repo (or --repo <path>). Artifacts go to <repo>/.temp/adk/implement/<task>/ (per shared/paths.md).
--detailed asks the skill to gather richer programmatic context before editing (for example, broader code-index queries or supporting docs). --deep asks the harness to use its stronger model profile per shared/model-depth.md. Auto-select --deep for ambiguous architecture choices, cross-module changes, migrations, auth/security/payments/permissions work, or large blast radius.
References (loaded as needed)
| Aspect | File |
|---|---|
| Input dispatch (which sub-flow) | references/dispatch.md |
| Workflow (Phase 0–4) | references/workflow.md |
Fork IDs (defaults trained by /adk-improve) |
references/forks.md |
| Hard rules + refusals | references/rules.md |
| Sub-flow detail | references/<sub-flow>.md — authored on first use of each sub-flow |
Cross-skill dependencies
- Edit-format:
shared/edit-format.md - Plan/Act mode:
shared/plan-act-mode.md - Model depth:
shared/model-depth.md - Constitution:
shared/constitution.md - Advisor + question-first:
shared/advisor.md,shared/question-first.md,shared/narration.md - Code index (TDD sub-flow):
shared/guidelines/code-index.md— when drafting a TDD from a Jira ticket, callfrom scripts.lib.code_index.query import open_index, similar, callersto surface related code patterns + blast-radius estimates. Fail open onIndexNotBuilt(prompt the user to runadk repo add <git-url>). Surfaceindexed_sha+age_daysin the draft.
Sub-flow status
| Sub-flow | Status |
|---|---|
| from-jira | spec is references/from-jira.md — authored on first real use |
| from-issue | spec is references/from-issue.md — authored on first real use |
| from-tdd / from-confluence / from-slack-thread / greenfield | authored on first real use |
The skill self-authors a sub-flow reference on its first invocation, asks the user to confirm, then saves it for future runs. This avoids filler content written without real signal.
References shipped
references/dispatch.md— see/adk-implement/references/dispatch.mdreferences/forks.md— see/adk-implement/references/forks.mdreferences/rules.md— see/adk-implement/references/rules.mdreferences/workflow.md— see/adk-implement/references/workflow.md