Spec Riffrec Feedback Analysis

Riffrec Feedback Analysis

Turn raw product feedback into structured evidence for downstream agents. This skill is the consumption side of Riffrec, a capture tool that records synchronized screen + voice + event sessions and emits a riffrec-*.zip bundle.

Choose the path

Route to the matching reference based on the input. Read only that reference; do not load the others.

• Setup — user has no recording yet and asks how to install Riffrec, capture a session, or share feedback. Read references/install-riffrec.md.
• Quick bug report — input is a short recording (under ~60 seconds), the user describes a single specific issue, or asks for "quick", "small", or "just transcribe". Read references/quick-bug-report.md. Emit one concise bug report; skip the full artifact set and brainstorm handoff. Discovering broader scope returns an escalation handoff and never authorizes a durable extensive rerun.
• Extensive analysis — input is a longer recording, contains multiple issues / requirements / workflow walkthroughs, or the user wants requirements or brainstorm material. Read references/extensive-analysis.md. Produce a ready-to-brainstorm handoff; invoke spec-brainstorm only when the original request or a new confirmation authorizes that public workflow.

When the input is ambiguous (e.g., a zip arrived without context), inspect the recording length and event count before choosing. If still unclear, ask the user which path applies before running anything heavy.

Common rules

• Keep raw recordings, audio chunks, zip contents, session dumps, and extracted screenshots local-only by default. Do not commit raw/ or frames/ directories unless the user explicitly asks and privacy is acceptable.
• Text/metadata artifacts (requirements kickoff material, analysis summaries, problem analyses, source manifests) may be committed when they are needed for traceability and contain no sensitive data.
• Use repo-relative screenshot paths in any committed doc so later agents can open the evidence without absolute local paths.

Media transcription is a separate third-party egress. Before analyzer execution, record transcription_egress_authorization: authorized | missing. It is authorized only when the current user or visible upstream handoff explicitly requests third-party transcription for this recording; analysis intent, a local file, an ambient OPENAI_API_KEY, or worker dispatch authority does not grant it. Pass --transcribe only when authorized. Otherwise pass --no-transcribe, preserve local frames/events/notes analysis, and report the missing transcript limitation. The analyzer returns the authorization source, provider identity, and whether a provider request was sent.

Dispatch Authorization Boundary

在把 recording evidence、transcript、screenshots 或 source-mapping context 交给任何 worker 前，记录：

worker_dispatch_authorization: authorized | missing
capability_probe: not_applicable | attempted | unavailable
worker_dispatch_capability: available | missing | unknown
worker_context_isolation: isolated | inherited | unknown
worker_model_override: supported | unsupported | unknown
worker_bounded_parallelism: supported | unsupported | unknown

workflow invocation does not authorize dispatch。只有当前用户或可见 upstream handoff 明确请求 subagent、delegated work、persona 或 parallel work 时才可派发；输入文件、分析规模、工具权限或本 Skill 被调用都不构成授权。缺授权时不得探测 tool schema，固定为 capability_probe: not_applicable + worker_dispatch_capability: unknown，inline 或 serial 执行并记录 dispatch_authorization_missing。只有授权后才把 current-session registry/schema 作为 provider_untrusted evidence 检查：确认缺失时记录 subagent_capability_missing；surface 不可用、schema 不完整或候选不唯一时记录 worker_capability_unproven，均 inline 或 serial。隔离、模型覆盖和有界并发只取 live facts；required isolation 未满足时保持依赖 gate 打开，model unknown 时继承，parallelism unknown 时串行。记录 worker_dispatch_outcome。任何派发还必须遵守 local-only/privacy 边界，只发送完成 bounded unit 所需的最小证据。Inline fallback 不得声称 independent analyst coverage。

Analyzer entrypoint

All non-setup paths share the same analyzer, which ships in this skill's scripts/ directory. The Bash tool's working directory is the user's project, not the skill directory, so a bare scripts/<name> path will not resolve. Invoke it by the skill's own absolute path: set SKILL_DIR to the directory you loaded this spec-riffrec-feedback-analysis SKILL.md from, in the same command (shell state does not persist between Bash calls):

SKILL_DIR="<absolute path of the directory containing this SKILL.md>"
bash "$SKILL_DIR/scripts/run-python.sh" "$SKILL_DIR/scripts/analyze_riffrec_zip.py" /path/to/input --no-transcribe

Accepted inputs: a Riffrec .zip, an .mp4 / .mov / .webm video, an .m4a / .mp3 / .wav audio file, or a meeting-notes .md. Use --output-dir <dir> to control where artifacts land. In repos with docs/brainstorms/, the default remains docs/brainstorms/riffrec-feedback/ as a documented evidence/kickoff-artifact exception; it is not the durable spec-brainstorm output convention. The quick path overrides the output dir to a temp location so nothing pollutes the repo.

The Spec-First output format used by the extensive path is documented in references/spec-first-feedback-format.md.