bc-registry-asset-comments-and-issues

name	bc-registry-asset-comments-and-issues
description	Use when academy staff need to manage registry asset comments (optional `academies=<id,id>` querystring on GET for multi-academy scope) and asset error issues end-to-end including deduping legacy duplicate error rows, and do NOT use for public asset browsing or non-registry workflows.
requires	[]

Skill: Manage Registry Asset Comments and Issues

When to Use

Use this skill when staff need to create, assign, review, resolve, update, or delete comments on registry assets, and when they need to list/update/delete asset error issues. Use it for academy-scoped moderation and QA workflows. Do NOT use this skill for public learner asset browsing or for editing syllabus/admissions entities outside registry.

Concepts

• An asset comment is a human note attached to an asset (text, assignment fields, and resolution flags).
• An asset issue here refers to an AssetErrorLog record with status lifecycle ERROR -> FIXED or IGNORED.
• An asset issue has a logical identity used for deduplication and grouped operations: the same slug + asset_type + path + asset (when asset is present). When asset is missing, the logical identity is slug + asset_type + path.
• Newly logged issues should collapse into that logical identity (re-logging updates the existing row instead of inserting another copy), and the database prevents multiple rows with the same identity going forward.
• A comment is considered "closed" by setting resolved=true (there is no separate comment status enum).
• Comments and error logs use different capabilities and different endpoints.
• Listing comments requires a single Academy header scope and supports optional academies=<id,id,...> on GET for read aggregation. The API returns only academies where the user has read_asset; if some requested academies are not allowed, response includes academy_scope metadata with requested/applied IDs and resolution=partial. If none are allowed, the API returns 403.

flowchart LR
    commentOpen[Comment resolved=false] -->|set resolved=true| commentResolved[Comment resolved=true]
    commentResolved -->|set resolved=false| commentOpen

flowchart LR
    errorState[ERROR] --> fixedState[FIXED]
    errorState --> ignoredState[IGNORED]
    fixedState --> errorState
    ignoredState --> errorState

Workflow

1. Set request headers for all calls:

   • Authorization: Bearer <token>
   • Academy: <academy_id> (required for /academy/ endpoints)
   • Optional: Accept-Language: en|es for translated errors.

2. For comment creation, resolve the target asset first:

   • Use the asset ID or slug known by the caller.
   • POST /v1/registry/academy/asset/comment accepts asset as either numeric ID or slug.

3. Create a comment:

   • Send asset and text.
   • author is set by the backend from the authenticated user.

4. Review/filter comments:

   • GET /v1/registry/academy/asset/comment with filters such as asset, resolved, delivered, owner, author, and academies.
   • The Academy header is required (single academy). For read aggregation, send academies=<id1,id2,...> in querystring; the API applies only academies where the user has read_asset, and partial application returns academy_scope metadata.
   • This list is paginated and sorted by newest (-created_at) by default.
   • Optional sort overrides are supported, including sort=priority and sort=-priority.

5. Manage comment lifecycle:

   • Use PUT /v1/registry/academy/asset/comment/<comment_id> to update assignment/flags.
   • Use DELETE /v1/registry/academy/asset/comment/<comment_id> to remove a comment.
   • To close a comment, set resolved=true.

6. Manage asset error issues:

   • List with GET /v1/registry/academy/asset/error.
   • Fetch error legend/catalog with GET /v1/registry/academy/asset/error/catalog.
   • If the UI still shows multiple identical rows for the same logical issue (legacy data), pick the row you want to keep and call POST /v1/registry/academy/asset/error/<error_id>/dedupe on that row id.
   • Update one or many with PUT /v1/registry/academy/asset/error (single object or list).
   • Delete one by URL ID or bulk-delete via query lookups.

7. Confirm outcomes after writes:

   • Re-fetch comment/error records after update/delete.
   • For error status updates, expect grouped status propagation across matching rows.

Endpoints

Asset Comments

Action	Method	Path	Required headers	Required body fields	Response notes
List comments	GET	/v1/registry/academy/asset/comment	Authorization, Academy	None	Paginated list (default sort: newest first). Supports sort=priority and sort=-priority. Use querystring academies=<id1,id2,...> for multi-academy read aggregation.
Create comment	POST	/v1/registry/academy/asset/comment	Authorization, Academy	asset, text	201 comment object; author set from session user.
Update comment	PUT	/v1/registry/academy/asset/comment/<comment_id>	Authorization, Academy	None globally required; send fields to change	200 updated comment object.
Delete comment	DELETE	/v1/registry/academy/asset/comment/<comment_id>	Authorization, Academy	None	204 no content.

Permissions

• GET requires read_asset.
• POST/PUT/DELETE require crud_asset.

List scope and filters

• academies=<id1,id2,...> on GET (optional multi-academy) — each ID must be numeric. The API applies only IDs where the user has read_asset on an ACTIVE academy. If some are filtered, response includes academy_scope metadata; if all are filtered, response is 403.
• asset=<id1,id2> or asset=<slug1,slug2> (mixed values are supported)
• resolved=true|false
• delivered=true|false
• owner=<owner_email>
• author=<author_email>
• sort=priority|-priority (optional; default remains -created_at)

Create comment request

{
  "asset": "javascript-arrays-intro",
  "text": "The instructions need clearer expected output examples."
}

Create comment response

{
  "id": 823,
  "text": "The instructions need clearer expected output examples.",
  "asset": {
    "id": 301,
    "slug": "javascript-arrays-intro",
    "title": "JavaScript Arrays Intro"
  },
  "resolved": false,
  "delivered": false,
  "priority": 0,
  "author": {
    "id": 17,
    "email": "[e-mail]"
  },
  "owner": null,
  "created_at": "2026-04-02T12:15:10.103Z"
}

Update comment request (close + assign)

{
  "owner": 42,
  "resolved": true,
  "delivered": true,
  "urgent": false,
  "priority": 2
}

Update comment response

{
  "id": 823,
  "text": "The instructions need clearer expected output examples.",
  "asset": {
    "id": 301,
    "slug": "javascript-arrays-intro",
    "title": "JavaScript Arrays Intro"
  },
  "resolved": true,
  "delivered": true,
  "priority": 2,
  "author": {
    "id": 17,
    "email": "[e-mail]"
  },
  "owner": {
    "id": 42,
    "email": "[e-mail]"
  },
  "created_at": "2026-04-02T12:15:10.103Z"
}

Asset Error Issues

Action	Method	Path	Required headers	Required body fields	Response notes
List issues	GET	/v1/registry/academy/asset/error	Authorization, Academy	None	Paginated list; supports issue filters and `sort=priority
Get issue catalog (legend)	GET	/v1/registry/academy/asset/error/catalog	Authorization, Academy	None	Returns dynamic catalog from AssetErrorLogType with descriptions and trigger hints.
Dedupe legacy duplicate issues	POST	/v1/registry/academy/asset/error/<error_id>/dedupe	Authorization, Academy	{} (empty JSON object)	200 JSON with kept, deleted_ids, deleted_count.
Update issues (single or bulk)	PUT	/v1/registry/academy/asset/error	Authorization, Academy	id per object, plus fields to update	200 list of updated issue objects.
Delete one issue	DELETE	/v1/registry/academy/asset/error/<error_id>	Authorization, Academy	None	204 no content.
Delete many issues	DELETE	/v1/registry/academy/asset/error?<lookups>	Authorization, Academy	Query lookups	204 no content.

Permissions

• GET requires read_asset_error.
• PUT/DELETE/POST dedupe require crud_asset_error.

Dedupe behavior (POST /v1/registry/academy/asset/error/<error_id>/dedupe)

• Call this on the issue row id you want to keep.
• The API deletes other rows that match the same logical identity as the kept row.
• If there are no duplicates, the API still returns 200 with deleted_count: 0.
• If duplicates exist, the API merges a small set of fields onto the kept row before deleting duplicates:
  • status uses precedence ERROR > FIXED > IGNORED across the kept row + duplicates.
  • priority becomes the maximum value across the kept row + duplicates.
  • If the kept row has an empty status_text, it is filled from duplicates (preferring the newest duplicate that has text).
  • If the kept row has no user, it may be filled from duplicates (preferring the newest duplicate that has a user).

Dedupe request

{}

Dedupe response (no duplicates)

{
  "kept": {
    "id": 911,
    "slug": "readme-syntax",
    "priority": 0,
    "status": "ERROR",
    "path": "README.md",
    "asset_type": "LESSON",
    "created_at": "2026-04-02T09:18:01.122Z"
  },
  "deleted_ids": [],
  "deleted_count": 0
}

Dedupe response (duplicates removed)

{
  "kept": {
    "id": 911,
    "slug": "readme-syntax",
    "priority": 3,
    "status": "ERROR",
    "path": "README.md",
    "asset_type": "LESSON",
    "created_at": "2026-04-02T09:18:01.122Z"
  },
  "deleted_ids": [912, 913],
  "deleted_count": 2
}

Catalog (legend) response fields

• slug: canonical error slug.
• label: short human-readable name.
• description: what the error means.
• common_trigger_situations: common scenarios where it is generated.
• severity_hint: triage hint (high, medium, low, unknown).
• status_notes: practical guidance for ERROR -> FIXED/IGNORED.

Catalog behavior

• The endpoint is dynamic and auto-discovers values from AssetErrorLogType.
• If a new error type is added in code, it appears automatically in the catalog.
• If a slug has no custom metadata yet, fallback defaults are returned so the endpoint remains stable.
• Supports multi-academy read aggregation through Academy header (for example, Academy: 1,2) with partial-scope metadata when some academies are not allowed.

List filters (GET /v1/registry/academy/asset/error)

• asset=<slug1,slug2>: exact asset slug match (asset__slug__in), lowercased server-side.
• slug=<slug1,slug2>: exact error slug match (slug__in), lowercased server-side.
• status=ERROR,FIXED,IGNORED: exact status match (status__in), uppercased server-side.
• asset_type=<type1,type2>: exact type match (asset_type__in), uppercased server-side.
• like=<search_text>: fuzzy match over error slug or path (icontains).
• sort=priority|-priority: optional ordering override (default remains newest first).

Filtering examples

• GET /v1/registry/academy/asset/error?status=ERROR&asset_type=PROJECT
• GET /v1/registry/academy/asset/error?slug=invalid-url,empty-readme
• GET /v1/registry/academy/asset/error?asset=javascript-arrays-intro,python-loops
• GET /v1/registry/academy/asset/error?like=telemetry

Update issue request (single)

{
  "id": 911,
  "status": "FIXED"
}

Update issue request (bulk list)

[
  {
    "id": 911,
    "status": "FIXED"
  },
  {
    "id": 912,
    "status": "IGNORED"
  }
]

Update issue response

[
  {
    "id": 911,
    "slug": "readme-syntax",
    "priority": 0,
    "status": "FIXED",
    "path": "README.md",
    "asset_type": "LESSON",
    "created_at": "2026-04-02T09:18:01.122Z"
  }
]

Edge Cases

1. Single or multiple academy IDs in header
   Academy can be a single ID (Academy: 1) or a comma-separated list (Academy: 1,2,3) for GET aggregation. Mutating endpoints still expect a single academy scope.

2. Comment text is immutable after creation
   PUT for comments cannot update text, asset, or author; only lifecycle/assignment fields should be changed.

3. Owner cannot self-resolve
   If the current session user is the same as owner, changing resolved raises a validation error.

4. Comment response does not expose all stored fields
   urgent exists on the model, but the comment response serializer does not return it.

5. Comment PUT status is not a supported lifecycle field
   The view checks status=NOT_STARTED, but comment serializer logic excludes author; do not rely on status when updating comments.

6. Error status updates can affect grouped rows
   Updating status for one error can propagate to other rows with the same slug, asset_type, path, and asset.

7. Error delete mode validation
   Do not mix URL error_id with bulk lookup query params in the same delete request.

8. Dedupe can change fields on the kept row
   If duplicates disagree on status, priority, status_text, or user, the merge rules on the dedupe endpoint may update the kept row before deleting duplicates. Re-fetch the kept id after dedupe if the UI caches the old values.

9. Recurring errors return to ERROR when re-logged
   If an issue was marked FIXED/IGNORED but the underlying problem happens again, the system may upsert the same logical row back to ERROR with refreshed details.

Checklist

1. Confirmed headers: Authorization and Academy are set on every /academy/ call.
2. Used comment POST with asset + text, and verified author came from session.
3. Used comment PUT only for assignment/lifecycle fields, not text.
4. Closed or reopened comments using resolved flag and handled owner-resolve restriction.
5. Listed and filtered comments/issues with proper query parameters and used multi-academy Academy header on GET when cross-academy aggregation was needed.
6. Updated error issues with awareness of grouped status propagation.
7. If duplicate identical issues were still visible, called POST /v1/registry/academy/asset/error/<keeper_id>/dedupe and verified deleted_count and the returned kept.id.
8. Deleted comments/issues using the correct single or bulk endpoint pattern.