Update to SUN
Snapshot Oct 29, 2026 · 19:12 UTC · version 1.1.3
Collection source: downloaded plugin package.
Instructions updated for status
Skill instructions — Now says: 1. Find the `job_id` in the result of the create tool earlier in this conversation. Never ask the user for it or show it to them: it is an internal...
Observed in instructions or declared skills. Runtime behavior has not been tested.
Up to 10 plugins free · no card required.
All fields and original values.
Skill instructions
Oct 26, 2026 · 06:08 UTC
1. Identify the `job_id`. If the conversation does not contain one, ask for it.
Oct 29, 2026 · 19:12 UTC
1. Find the `job_id` in the result of the create tool earlier in this conversation. Never ask the user for it or show it to them: it is an internal identifier they never see. If the conversation has none, say you can only check ...
Compare saved observations
Download comparison JSONSKILL.md text changes
− removed · + added
--- before +++ after @@ -6,7 +6,10 @@ Use this skill when someone asks about a podcast, meditation, story or song that is already generating or wants the finished audio. -1. Identify the `job_id`. If the conversation does not contain one, ask for it. +1. Find the `job_id` in the result of the create tool earlier in this + conversation. Never ask the user for it or show it to them: it is an + internal identifier they never see. If the conversation has none, say you + can only check a generation started in this conversation. 2. Call `get_course_status`. It reports `PENDING`, `PROCESSING`, `SUCCESS`, or `ERROR`, and a `progress` value from 0 to 100 once available. 3. While it is `PENDING` or `PROCESSING` and the user wants to wait, call
Full technical diff · 1 changed fields
changed /skill_md_contents
"---\nname: status\ndescription: Check the progress of a podcast, meditation, story or song generated by SUN and fetch it when ready. Use when someone asks whether it is done, how far along it is, or for the finished audio.\n---\n\nUse this skill when someone asks about a podcast, meditation, story or song that\nis already generating or wants the finished audio.\n\n1. Identify the `job_id`. If the conversation does not contain one, ask for it.\n2. Call `get_course_status`. It reports `PENDING`, `PROCESSING`, `SUCCESS`, or\n `ERROR`, and a `progress` value from 0 to 100 once available.\n3. While it is `PENDING` or `PROCESSING` and the user wants to wait, call\n `wait_for_course` instead of polling yourself, and again each time it\n returns `state: waiting`; a 30-minute podcast typically takes 60 to 300\n seconds before its first audio.\n4. Once it reaches `SUCCESS` (or `wait_for_course` returns `state: playable`),\n call `get_course` and report the name with its `listen_url` as a clickable\n link. Describe the result by its `content_type`:\n - `course`: present the returned `lectures` as episodes with their audio\n links.\n - `meditation`: one guided session, even though it arrives in several parts.\n - `story`: its `lectures` are the chapters.\n - `song`: share the `lyrics` when present; a song without lyrics is\n instrumental.\n5. If the user is impatient, call `get_course` with `include_partial` set to\n `true` to surface parts that have already finished; its `listen_url` plays\n them, and is null while nothing can play yet.\n\nAudio links expire after seven days. Say so when sharing them, and never treat\nan audio link as a durable identifier. `listen_url` does not expire.\n\nOn `ERROR`, report the error the tool returns. Do not retry automatically and\ndo not start a new generation without asking, since that spends the user's\ncredits.\n"
"---\nname: status\ndescription: Check the progress of a podcast, meditation, story or song generated by SUN and fetch it when ready. Use when someone asks whether it is done, how far along it is, or for the finished audio.\n---\n\nUse this skill when someone asks about a podcast, meditation, story or song that\nis already generating or wants the finished audio.\n\n1. Find the `job_id` in the result of the create tool earlier in this\n conversation. Never ask the user for it or show it to them: it is an\n internal identifier they never see. If the conversation has none, say you\n can only check a generation started in this conversation.\n2. Call `get_course_status`. It reports `PENDING`, `PROCESSING`, `SUCCESS`, or\n `ERROR`, and a `progress` value from 0 to 100 once available.\n3. While it is `PENDING` or `PROCESSING` and the user wants to wait, call\n `wait_for_course` instead of polling yourself, and again each time it\n returns `state: waiting`; a 30-minute podcast typically takes 60 to 300\n seconds before its first audio.\n4. Once it reaches `SUCCESS` (or `wait_for_course` returns `state: playable`),\n call `get_course` and report the name with its `listen_url` as a clickable\n link. Describe the result by its `content_type`:\n - `course`: present the returned `lectures` as episodes with their audio\n links.\n - `meditation`: one guided session, even though it arrives in several parts.\n - `story`: its `lectures` are the chapters.\n - `song`: share the `lyrics` when present; a song without lyrics is\n instrumental.\n5. If the user is impatient, call `get_course` with `include_partial` set to\n `true` to surface parts that have already finished; its `listen_url` plays\n them, and is null while nothing can play yet.\n\nAudio links expire after seven days. Say so when sharing them, and never treat\nan audio link as a durable identifier. `listen_url` does not expire.\n\nOn `ERROR`, report the error the tool returns. Do not retry automatically and\ndo not start a new generation without asking, since that spends the user's\ncredits.\n"
Full snapshot data
{
"description": "Check the progress of a podcast, meditation, story or song generated by SUN and fetch it when ready. Use when someone asks whether it is done, how far along it is, or for the finished audio.",
"included_files": [
{
"relative_path": "agents/openai.yaml",
"size_in_bytes": 435
}
],
"name": "status",
"skill_md_contents": "---\nname: status\ndescription: Check the progress of a podcast, meditation, story or song generated by SUN and fetch it when ready. Use when someone asks whether it is done, how far along it is, or for the finished audio.\n---\n\nUse this skill when someone asks about a podcast, meditation, story or song that\nis already generating or wants the finished audio.\n\n1. Find the `job_id` in the result of the create tool earlier in this\n conversation. Never ask the user for it or show it to them: it is an\n internal identifier they never see. If the conversation has none, say you\n can only check a generation started in this conversation.\n2. Call `get_course_status`. It reports `PENDING`, `PROCESSING`, `SUCCESS`, or\n `ERROR`, and a `progress` value from 0 to 100 once available.\n3. While it is `PENDING` or `PROCESSING` and the user wants to wait, call\n `wait_for_course` instead of polling yourself, and again each time it\n returns `state: waiting`; a 30-minute podcast typically takes 60 to 300\n seconds before its first audio.\n4. Once it reaches `SUCCESS` (or `wait_for_course` returns `state: playable`),\n call `get_course` and report the name with its `listen_url` as a clickable\n link. Describe the result by its `content_type`:\n - `course`: present the returned `lectures` as episodes with their audio\n links.\n - `meditation`: one guided session, even though it arrives in several parts.\n - `story`: its `lectures` are the chapters.\n - `song`: share the `lyrics` when present; a song without lyrics is\n instrumental.\n5. If the user is impatient, call `get_course` with `include_partial` set to\n `true` to surface parts that have already finished; its `listen_url` plays\n them, and is null while nothing can play yet.\n\nAudio links expire after seven days. Say so when sharing them, and never treat\nan audio link as a durable identifier. `listen_url` does not expire.\n\nOn `ERROR`, report the error the tool returns. Do not retry automatically and\ndo not start a new generation without asking, since that spends the user's\ncredits.\n"
}SHA-256 of public snapshot: e5514faa28c576da0bd85b314ab3a61a6c440451361649aa01aa3f3bdfd41035