菜单
Log a corrective interaction — a command produced something wrong and you're fixing it — as plugin feedback, then act on your correction directly. Captures the friction, your verbatim prompt, and the resolution to the specs repo for the maintainer. Activated when the user prompt starts with "prompt:".
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
Jira-driven release-notes drafting. Reads a Value Increment (or any ticket) from exported markdown, optionally grounds in PR diffs, renders a dynatrace-docs release-notes body, runs a light dt-style-checker gate, and writes a persistent draft to paste into Jira's release-notes field. Activated when the user prompt starts with "release-notes:".
Architecture-authoring workflow (Product Architect phase, sub-project 3 of the VI-creation flow). Grounds on the mounted implementation repos (architect-driven discovery — no PRs) and authors an ARD for a VI (create-ard: <VI-KEY>) or an Epic (create-ard: <VI-KEY> <Epic-KEY>, inheriting the VI-level ARD), against _shared/ard-format.md, gated by the Opus ard-reviewer, written to $SPECS_PATH/specifications/<KEY>-<slug>/. Optional; scoped; product-architecture level (no code writing). Introduces the pa role. Activated when the user prompt starts with "create-ard:".
VI-creation workflow (PM phase, sub-project 2 of the VI-creation flow). Turns a refined idea.md + a user-supplied JIRA-KEY into a high-quality Value Increment document (spine + adapt-in profiles: --lean|--hybrid|--full), authored via a relentless grill against _shared/vi-format.md, gated by the Opus vi-reviewer, written to $SPECS_PATH/specifications/<KEY>-<slug>/ and published to Jira by paste + re-import. Product-level (no code scan). Offers release-notes: and create-ard: as next steps. Activated when the user prompt starts with "create-vi:".
VI-update workflow (PM phase) — refresh/re-do an existing Value Increment. Resolves the VI Jira-import-first (source of truth) with a 3-day freshness gate, grounds on the VI + comments + any ARD/spec/transcript, updates it via a relentless grill against _shared/vi-format.md, gated by the Opus vi-reviewer, and writes canonical + archived revisions to $SPECS_PATH/specifications/<KEY>-<slug>/. Product-level (no code scan). Activated when the user prompt starts with "update-vi:".
Jira-driven feature-documentation workflow. Phase 0 preflight-discovers the docs repo + profile (in-repo → built-in dynatrace-docs default → on-demand docs-profile:) and the VI's specs dir under /workspace. Phase 4.5 determines/confirms the applicable space(s). Optional saas|managed constraint scopes the run to one space. Reads a Value Increment hierarchy from exported markdown, resolves PR diffs in parallel, synthesises product documentation, and gates on style-check and Opus doc review. Optional --counterpart <JiraID|PR-url> grounds a space-constrained run on the other space's existing docs (read-only). Activated when the user prompt starts with "document:".
Scan the vault (or a target directory) for tags used in page frontmatter and page bodies, diff against the vault's .obsidian/copilot/tag-index.md, prompt the user to approve new tags and clean stale ones, then update tag-index.md in place. Run after heavy ingest or other large content changes when new concepts may have introduced new tags. Triggers on: wiki-tags-refresh, refresh tags, update tag index, wiki tags, sync tags.
| name | prompt |
| description | Log a corrective interaction — a command produced something wrong and you're fixing it — as plugin feedback, then act on your correction directly. Captures the friction, your verbatim prompt, and the resolution to the specs repo for the maintainer. Activated when the user prompt starts with "prompt:". |
| allowed-tools | view, edit, create, bash, glob, grep, task, ask_user |
Log a corrective interaction and act on it: the argument (text following the prompt: trigger)
prompt: is for when a dev-workflows command (specify:, design:,
implement:, document:, …) produced something wrong and you want to correct
it directly. It captures the corrective triple as plugin feedback, then
performs the correction. origin: prompt.
Captured (per ~/.copilot/installed-plugins/ihudak-copilot-plugins/dev-workflows/skills/_shared/feedback-emission.md §1):
prompt: argument).Infer the target command from recent context — which command's output you are
correcting. Ask only if genuinely ambiguous (one grouped prompt, last choice
"Other… (describe)"). If no command applies, use n/a.
Perform the corrective request in the prompt: argument directly (the quick correction).
Keep a one-line summary of what you did — this becomes the Resolution block.
Cite ~/.copilot/installed-plugins/ihudak-copilot-plugins/dev-workflows/skills/_shared/feedback-emission.md and call its
emit-prompt entry point (§6). Provide:
prompt: argument, verbatim (never paraphrased).command (Phase 1), an inferred category (§1 vocab, reuse-first), impact,
jira_key (or null), source.emit-prompt resolves the write target via the §2 specs-first ladder, formats
the entry with the two extra prose blocks (origin: prompt), and appends per §3
(prompt entries are never silently skipped). Write silently — a single append.
Surface the persisted path and any degradation notice. This command NEVER commits, and NEVER writes into a docs/code repo or the current working directory (only the correction itself edits your target files, as you requested).