메뉴
Transcript generator for CoQuill. Reads an interview_log.json and manifest.yaml, then writes a human-readable transcript.md to the job folder. Called by the coquill orchestrator — not triggered directly by the user.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Template analyzer for CoQuill (v2). Parses docx/HTML templates, extracts variables including conditionals and loops, merges config.yaml overrides, infers types, and generates a v2 manifest.yaml. Called by the coquill orchestrator — not triggered directly by the user.
Document renderer for CoQuill. Takes a template, variable values, and produces rendered documents (docx or html+pdf). Validates output for unfilled placeholders. Called by the coquill orchestrator — not triggered directly by the user.
Document assembly tool. Matches user requests to docx/HTML templates, interviews the user for variable values, and renders completed documents. Supports conditional sections, loops, and developer-configured interview flows. Trigger when the user says: 'prepare a document', 'draft a [template name]', 'fill out a template', 'I need an NDA/contract/agreement', or any request that implies assembling a document from a template.
| name | coquill-transcriber |
| description | Transcript generator for CoQuill. Reads an interview_log.json and manifest.yaml, then writes a human-readable transcript.md to the job folder. Called by the coquill orchestrator — not triggered directly by the user. |
You generate a transcript from a completed interview session. The heavy lifting
is handled by scripts/transcribe.py; your job is to invoke it and relay the result.
interview_log_path — path to interview_log.json in the job foldermanifest_path — path to manifest.yaml in the template directoryjob_folder — path to the job output folderoutput_files — comma-separated output file basenames (e.g., agreement.docx, agreement.pdf)ended_at — ISO 8601 timestamp for when the document was renderedThe interview log records the substance of each exchange, not a literal word-for-word
transcript of every micro-turn. When a user answered multiple questions at once, the log
records the net result. The clarification entry handles the exceptional case where
the user asked a substantive question about the document before answering.
Resolve the script path relative to the project root and invoke it:
python scripts/transcribe.py \
--interview-log <interview_log_path> \
--manifest <manifest_path> \
--job-folder <job_folder> \
--output-files "<output_files>" \
--ended-at "<ended_at>"
The script reads both files, builds the four transcript sections (header, interview,
confirmed values, footer), writes transcript.md to the job folder, and prints a
JSON result to stdout.
The script prints JSON: {"transcript_path": "...", "success": true} on success,
or {"transcript_path": null, "success": false, "error": "..."} on failure.