From junior-architect-position-plugin
Chunk-read long transcripts against specific questions, returning traceable direct quotes (discovery: rubber-duck). Evaluate when task requires reading meeting transcripts, interview recordings, or conversation logs from Google Drive or Fireflies. Triggers: 'read transcript', 'extract from transcript', 'what did they say about', 'find quotes from meeting'.
npx claudepluginhub mariuswilsch/developer-positionThis skill uses the workspace's default tool permissions.
Read long transcripts methodically by chunking, then extracting question-relevant quotes with source traceability. Prevents the common failure mode of search+grep on transcripts, which loses synthesis quality.
Guides Next.js Cache Components and Partial Prerendering (PPR) with cacheComponents enabled. Implements 'use cache', cacheLife(), cacheTag(), revalidateTag(), static/dynamic optimization, and cache debugging.
Migrates code, prompts, and API calls from Claude Sonnet 4.0/4.5 or Opus 4.1 to Opus 4.5, updating model strings on Anthropic, AWS, GCP, Azure platforms.
Performs token-optimized structural code search using tree-sitter AST parsing to discover symbols, outline files, and unfold code without reading full files.
Read long transcripts methodically by chunking, then extracting question-relevant quotes with source traceability. Prevents the common failure mode of search+grep on transcripts, which loses synthesis quality.
Google Drive transcript:
mcp-cli call google_workspace/get_drive_file_content '{"file_id": "<FILE_ID>"}'
Save output to /tmp/transcript.txt.
Fireflies transcript:
mcp-cli call hand-picked-tools/fireflies__fireflies_get_transcript '{"transcript_id": "<ID>"}'
Save output to /tmp/transcript.txt.
If source type is unclear, ask the user.
uv run python "$([ -f "${CLAUDE_PLUGIN_ROOT}/lib/chunk_transcript.py" ] && echo "${CLAUDE_PLUGIN_ROOT}/lib/chunk_transcript.py" || echo ~/.claude/lib/chunk_transcript.py)" /tmp/transcript.txt
Output: numbered files in /tmp/ (e.g., transcript_chunk1.txt, transcript_chunk2.txt).
Read each chunk file in its entirety. That is the purpose of chunking — do NOT search or grep within chunks.
For EACH chunk, read the full content and extract:
[chunk#N] where N is the chunk numberFormat each extracted quote:
[chunk#N] "Direct quote from transcript" — Speaker (if known)
When multiple questions are provided, organize extraction per-question within each chunk. Do NOT skip chunks — even chunks with no relevant quotes confirm absence.
Group all extracted quotes by question:
## Q: [Question text]
1. [chunk#2] "Relevant quote here" — Speaker
2. [chunk#5] "Another relevant quote" — Speaker
3. [chunk#5] "Follow-up point" — Speaker
## Q: [Second question]
1. [chunk#1] "Quote addressing this question" — Speaker
Include a summary synthesis after the quotes for each question, noting which chunks contained the most relevant material.
| Source | Identifier Pattern | MCP Tool |
|---|---|---|
| Google Drive | URL with /d/{fileId}/ or file ID | google_workspace/get_drive_file_content |
| Fireflies | Transcript ID or meeting title | hand-picked-tools/fireflies__fireflies_search → fireflies_get_transcript |
When the user provides a URL, extract the file ID. When they provide a meeting name, search Fireflies first.