peon-ping-remix

# peon-ping-remix

You are executing ONE reroll job for a draft PeonPing pack. Work only inside
the job's `draft_dir`. Do not touch any other directory, do not install
anything, do not commit anything.

## If invoked without a job-file path

This skill is internal: it is only ever invoked headlessly by the peon eval
server with a job-file path. If you were invoked WITHOUT one (a human ran
this directly), do nothing except tell them this skill runs automatically
from the eval server and that they should run `peon eval <pack>` instead.
Do not touch any files. Exit.

## Procedure

1. Read the job JSON at the path you were given: `scope` ("sound" or
   "pack"), `category`, `index`, `caption`, `draft_dir`.
2. Read `<draft_dir>/openpeon.json` and `<draft_dir>/prompts.json`. If
   `prompts.json` is missing, fail with a clear message — you cannot reroll
   without the current prompts.
3. Determine the target sound files: scope "sound" → the one file at
   `categories[category].sounds[index].file`; scope "pack" → every file in
   every category.
4. For each target, author a NEW render input honoring the caption:
   - Keep the pack's established aesthetic (read the other prompts for
     context). The caption is a correction, not a reset: "too harsh, want
     softer" means adjust intensity, keep the instrument palette unless the
     caption says otherwise.
   - sfx entries: rewrite `prompt`. tts entries: rewrite `text` (keep
     `voice_id` — never change a pack's voice in a reroll).
   - If the caption is empty, produce a fresh variation of the same idea.
5. Render each target: write the new input as a job file **under
   `<draft_dir>/jobs/`** (e.g. `<draft_dir>/jobs/render-job-<category>_<index>.json`
   — `approve` prunes everything under `jobs/`, so a render input left at the
   draft root would ship as junk in the approved pack) and run
   `python3 <peon-ping>/scripts/pack-render.py --job <file>` with `out` set
   to the target WAV path (overwrite in place). The renderer exits nonzero
   on silent/failed renders — if it fails, STOP and exit nonzero yourself
   with its stderr; do not half-update the log.
   (Resolve `<peon-ping>` as the scripts directory next to the running peon
   install, in order: `$PEON_DIR/scripts` when set, else
   `${CLAUDE_CONFIG_DIR:-$HOME/.claude}/hooks/peon-ping/scripts`, else the
   repo checkout you were invoked from.)
6. After ALL targets render successfully: update `prompts.json` with the new
   inputs, and append one entry per rendered sound to
   `<draft_dir>/eval-log.json` (create as `[]` if missing):
   `{"ts": "<ISO8601>", "scope", "category", "index", "caption",
     "old_prompt", "new_prompt", "file"}`.
7. Print a one-line summary per rendered sound. Exit 0.

## Hard rules

- NEVER write outside `draft_dir`.
- NEVER print or log the ElevenLabs key.
- A failed render leaves the previous WAV in place — pack-render.py renders
  into a same-directory temp file and only replaces the out path (atomic
  `os.replace`) on a successful, non-silent render; do not delete a WAV
  before rendering.
- Captions are the user's judgment. Honor them literally before creatively.