Fix format_ids schema to use structured format ID objects - #123
Merged
Conversation
Changed format_ids from arrays of strings to arrays of format-id objects (with agent_url and id fields) across all schemas and documentation. **Schema changes:** - static/schemas/v1/media-buy/package-request.json - static/schemas/v1/media-buy/list-creative-formats-request.json - static/schemas/v1/creative/list-creative-formats-request.json - static/schemas/v1/media-buy/get-products-request.json **Documentation updates (16 examples):** - docs/media-buy/task-reference/create_media_buy.md (10 examples) - docs/media-buy/product-discovery/media-products.md (3 examples) - docs/media-buy/advanced-topics/pricing-models.md (2 examples) - docs/media-buy/task-reference/list_creative_formats.md (1 example) This aligns with the protocol's requirement that format IDs are ALWAYS structured objects to avoid parsing ambiguity and handle namespace collisions between creative agents. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
The schema correctly uses format_ids consistently throughout. Updated the 'Better Field Naming' example to reflect the actual field name used. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
bokelley
added a commit
to prebid/salesagent
that referenced
this pull request
Oct 15, 2025
- Downloaded latest package-request.json schema (now uses FormatId objects) - Regenerated all Pydantic schemas from latest AdCP spec - Fixed schema generator bug: format_ids now correctly uses list[FormatId] not list[str] - Updated server comment to reflect new spec (both request and response use FormatId) - Maintains backward compatibility with string format IDs from older clients Per adcontextprotocol/adcp#123: - format_ids changed from array of strings to array of FormatId objects - FormatId structure: {agent_url: string, id: string} - Avoids parsing ambiguity and handles namespace collisions
bokelley
added a commit
to prebid/salesagent
that referenced
this pull request
Oct 15, 2025
* Fix format_ids handling: separate request and response fields
**Problem:**
The Package model had a validator that automatically converted format_ids
(array of strings) to format_ids_to_provide (array of FormatId objects)
during model initialization. This caused validation errors when clients
sent v1.8.0 requests with format_ids as strings, because the validator
ran too early and converted the data before the server could process it.
**Root Cause:**
- AdCP v1.8.0 spec defines TWO different fields:
- `format_ids` (request): array of strings in package-request.json
- `format_ids_to_provide` (response): array of FormatId objects in core/package.json
- The Package model was trying to serve both purposes with automatic migration
- The @model_validator ran during request deserialization, breaking client compatibility
**Solution:**
1. **schemas.py**: Removed the automatic migration validator
- Added explicit `format_ids: list[str]` field for requests
- Kept `format_ids_to_provide: list[FormatId]` field for responses
- Package model now accepts both formats without auto-conversion
2. **main.py**: Added explicit conversion in create_media_buy response building
- Convert format_ids → format_ids_to_provide when building response
- Happens after request validation, before response serialization
- Removes format_ids from response (only format_ids_to_provide in output)
**Testing:**
- All existing package-related unit tests pass (20 tests)
- Manual testing confirms Package model accepts format_ids as strings
- No breaking changes to existing code paths
**Result:**
✅ Clients can now send format_ids: ["display_300x250"] (v1.8.0 format)
✅ Server accepts request without validation errors
✅ Server responds with format_ids_to_provide: [{"agent_url": "...", "id": "..."}]
✅ Full AdCP v1.8.0 compatibility maintained
* Update schemas to use FormatId objects per AdCP PR #123
- Downloaded latest package-request.json schema (now uses FormatId objects)
- Regenerated all Pydantic schemas from latest AdCP spec
- Fixed schema generator bug: format_ids now correctly uses list[FormatId] not list[str]
- Updated server comment to reflect new spec (both request and response use FormatId)
- Maintains backward compatibility with string format IDs from older clients
Per adcontextprotocol/adcp#123:
- format_ids changed from array of strings to array of FormatId objects
- FormatId structure: {agent_url: string, id: string}
- Avoids parsing ambiguity and handles namespace collisions
* Fix schema generator: regenerate with correct package-request schema
Root cause: Outdated cached schema file had format_ids as list[str] instead of
list[FormatId] with $ref to format-id.json.
Investigation findings:
1. Schema generator's resolve_refs_in_schema() function works correctly
2. Issue was outdated cached file: tests/e2e/schemas/v1/_schemas_v1_media-buy_package-request_json.json
3. Old cached schema had: items: {type: 'string'}
4. New cached schema has: items: {$ref: '/schemas/v1/core/format-id.json'}
5. After downloading latest schema, generator correctly resolved $ref to FormatId
Result:
✅ Generated schemas now correctly use list[FormatId] (lines 378-384, 712-718)
✅ FormatId structure properly inlined (lines 13-29) with agent_url and id fields
✅ Schema generator working as designed - just needed updated input files
Lesson: Always ensure cached schemas are up-to-date before regenerating.
* Enhance schema sync checker to validate format_ids structure
Adds automated validation to prevent outdated schemas from causing
schema generator issues in the future.
Changes:
1. Added package-request.json and format-id.json to critical schemas list
2. Added new validate_format_ids_structure() check that specifically validates:
- format_ids field uses $ref to /schemas/v1/core/format-id.json
- Detects if format_ids incorrectly uses {type: 'string'}
- Provides clear error messages with fix instructions
3. Runs as first check in pre-commit hook for early detection
Prevention:
This prevents the root cause issue from PR investigation:
- Old cached schema had format_ids: {items: {type: 'string'}}
- New spec requires format_ids: {items: {$ref: '/schemas/v1/core/format-id.json'}}
- Schema generator was working correctly, but had wrong input
Testing:
✅ Runs successfully: uv run python scripts/check_schema_sync.py
✅ Validates format_ids structure correctly
✅ Provides actionable error messages
Result:
Future schema updates will be caught by pre-commit hook before
causing schema generator issues.
* Add comprehensive schema update documentation
Documents the complete schema update process to prevent future issues.
Contents:
- Overview of schema caching and validation
- Automated protection via pre-commit hooks
- Step-by-step update procedures (automatic and manual)
- Troubleshooting guide for common issues
- Schema validation checks explanation
- Best practices and workflow diagrams
- Reference to AdCP PR #123 format_ids changes
Key sections:
1. When to update schemas
2. How to update (automatic vs manual)
3. Schema validation checks (format_ids, critical schemas, full sync)
4. Troubleshooting common issues
5. Update workflow diagram
6. Best practices
This prevents the issue we encountered where outdated cached schemas
caused the generator to produce incorrect types.
* Sync all schemas and regenerate with latest AdCP spec
Systematic fix to ensure all cached schemas match live registry.
Changes:
1. Updated cached schemas (tests/e2e/schemas/v1/):
- get-products-request.json
- list-creative-formats-request.json
2. Regenerated all Pydantic models (src/core/schemas_generated/):
- 77 schema files updated
- All models now use latest schema definitions
- format_ids correctly uses list[FormatId] throughout
Process:
- Ran: uv run python scripts/check_schema_sync.py --update
- Verified: All schemas now in sync with adcontextprotocol.org
- Regenerated: uv run python scripts/generate_schemas.py
Validation:
✅ All schemas in sync
✅ format_ids structure validated
✅ Schema generator working correctly
✅ No errors in schema sync check
This ensures the codebase is fully synchronized with the official
AdCP spec and prevents future schema drift issues.
* Update documentation: clarify automated schema validation
Enhanced schema-updates.md to clearly explain automated protection.
Changes:
- Added 'Pre-commit Hook (Runs Automatically)' section
- Specified that validation runs on EVERY commit
- Listed exactly what gets validated (format IDs, 7 critical schemas, 30+ endpoints)
- Clarified that commits are BLOCKED if schemas out of sync
- Added clear error message example with fix instructions
This makes it crystal clear that the automation is:
1. Always running (every commit)
2. Comprehensive (all schemas checked)
3. Enforced (commits blocked on failure)
4. Actionable (clear fix instructions)
Result: Developers understand the protection is automatic and systematic.
* Fix schema adapter to handle format_ids conversion
- Handle both dict and Pydantic model cases for filters in GetProductsRequest.to_generated()
- Update test expectations to reflect AdCP PR #123 (format_ids as FormatId objects)
- Tests now pass: schema adapters properly convert string format_ids to FormatId objects during roundtrip
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
---------
Co-authored-by: Claude <noreply@anthropic.com>
bokelley
added a commit
that referenced
this pull request
Oct 15, 2025
* Fix format_ids schema to use structured format ID objects Changed format_ids from arrays of strings to arrays of format-id objects (with agent_url and id fields) across all schemas and documentation. **Schema changes:** - static/schemas/v1/media-buy/package-request.json - static/schemas/v1/media-buy/list-creative-formats-request.json - static/schemas/v1/creative/list-creative-formats-request.json - static/schemas/v1/media-buy/get-products-request.json **Documentation updates (16 examples):** - docs/media-buy/task-reference/create_media_buy.md (10 examples) - docs/media-buy/product-discovery/media-products.md (3 examples) - docs/media-buy/advanced-topics/pricing-models.md (2 examples) - docs/media-buy/task-reference/list_creative_formats.md (1 example) This aligns with the protocol's requirement that format IDs are ALWAYS structured objects to avoid parsing ambiguity and handle namespace collisions between creative agents. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * Remove outdated formats_to_provide reference from CLAUDE.md The schema correctly uses format_ids consistently throughout. Updated the 'Better Field Naming' example to reflect the actual field name used. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> --------- Co-authored-by: Claude <noreply@anthropic.com>
bokelley
added a commit
that referenced
this pull request
Oct 15, 2025
* Fix format_ids schema to use structured format ID objects Changed format_ids from arrays of strings to arrays of format-id objects (with agent_url and id fields) across all schemas and documentation. **Schema changes:** - static/schemas/v1/media-buy/package-request.json - static/schemas/v1/media-buy/list-creative-formats-request.json - static/schemas/v1/creative/list-creative-formats-request.json - static/schemas/v1/media-buy/get-products-request.json **Documentation updates (16 examples):** - docs/media-buy/task-reference/create_media_buy.md (10 examples) - docs/media-buy/product-discovery/media-products.md (3 examples) - docs/media-buy/advanced-topics/pricing-models.md (2 examples) - docs/media-buy/task-reference/list_creative_formats.md (1 example) This aligns with the protocol's requirement that format IDs are ALWAYS structured objects to avoid parsing ambiguity and handle namespace collisions between creative agents. 🤖 Generated with [Claude Code](https://claude.com/claude-code) * Remove outdated formats_to_provide reference from CLAUDE.md The schema correctly uses format_ids consistently throughout. Updated the 'Better Field Naming' example to reflect the actual field name used. 🤖 Generated with [Claude Code](https://claude.com/claude-code) --------- Co-authored-by: Claude <noreply@anthropic.com>
danf-newton
pushed a commit
to Newton-Research-Inc/salesagent
that referenced
this pull request
Nov 24, 2025
* Fix format_ids handling: separate request and response fields
**Problem:**
The Package model had a validator that automatically converted format_ids
(array of strings) to format_ids_to_provide (array of FormatId objects)
during model initialization. This caused validation errors when clients
sent v1.8.0 requests with format_ids as strings, because the validator
ran too early and converted the data before the server could process it.
**Root Cause:**
- AdCP v1.8.0 spec defines TWO different fields:
- `format_ids` (request): array of strings in package-request.json
- `format_ids_to_provide` (response): array of FormatId objects in core/package.json
- The Package model was trying to serve both purposes with automatic migration
- The @model_validator ran during request deserialization, breaking client compatibility
**Solution:**
1. **schemas.py**: Removed the automatic migration validator
- Added explicit `format_ids: list[str]` field for requests
- Kept `format_ids_to_provide: list[FormatId]` field for responses
- Package model now accepts both formats without auto-conversion
2. **main.py**: Added explicit conversion in create_media_buy response building
- Convert format_ids → format_ids_to_provide when building response
- Happens after request validation, before response serialization
- Removes format_ids from response (only format_ids_to_provide in output)
**Testing:**
- All existing package-related unit tests pass (20 tests)
- Manual testing confirms Package model accepts format_ids as strings
- No breaking changes to existing code paths
**Result:**
✅ Clients can now send format_ids: ["display_300x250"] (v1.8.0 format)
✅ Server accepts request without validation errors
✅ Server responds with format_ids_to_provide: [{"agent_url": "...", "id": "..."}]
✅ Full AdCP v1.8.0 compatibility maintained
* Update schemas to use FormatId objects per AdCP PR prebid#123
- Downloaded latest package-request.json schema (now uses FormatId objects)
- Regenerated all Pydantic schemas from latest AdCP spec
- Fixed schema generator bug: format_ids now correctly uses list[FormatId] not list[str]
- Updated server comment to reflect new spec (both request and response use FormatId)
- Maintains backward compatibility with string format IDs from older clients
Per adcontextprotocol/adcp#123:
- format_ids changed from array of strings to array of FormatId objects
- FormatId structure: {agent_url: string, id: string}
- Avoids parsing ambiguity and handles namespace collisions
* Fix schema generator: regenerate with correct package-request schema
Root cause: Outdated cached schema file had format_ids as list[str] instead of
list[FormatId] with $ref to format-id.json.
Investigation findings:
1. Schema generator's resolve_refs_in_schema() function works correctly
2. Issue was outdated cached file: tests/e2e/schemas/v1/_schemas_v1_media-buy_package-request_json.json
3. Old cached schema had: items: {type: 'string'}
4. New cached schema has: items: {$ref: '/schemas/v1/core/format-id.json'}
5. After downloading latest schema, generator correctly resolved $ref to FormatId
Result:
✅ Generated schemas now correctly use list[FormatId] (lines 378-384, 712-718)
✅ FormatId structure properly inlined (lines 13-29) with agent_url and id fields
✅ Schema generator working as designed - just needed updated input files
Lesson: Always ensure cached schemas are up-to-date before regenerating.
* Enhance schema sync checker to validate format_ids structure
Adds automated validation to prevent outdated schemas from causing
schema generator issues in the future.
Changes:
1. Added package-request.json and format-id.json to critical schemas list
2. Added new validate_format_ids_structure() check that specifically validates:
- format_ids field uses $ref to /schemas/v1/core/format-id.json
- Detects if format_ids incorrectly uses {type: 'string'}
- Provides clear error messages with fix instructions
3. Runs as first check in pre-commit hook for early detection
Prevention:
This prevents the root cause issue from PR investigation:
- Old cached schema had format_ids: {items: {type: 'string'}}
- New spec requires format_ids: {items: {$ref: '/schemas/v1/core/format-id.json'}}
- Schema generator was working correctly, but had wrong input
Testing:
✅ Runs successfully: uv run python scripts/check_schema_sync.py
✅ Validates format_ids structure correctly
✅ Provides actionable error messages
Result:
Future schema updates will be caught by pre-commit hook before
causing schema generator issues.
* Add comprehensive schema update documentation
Documents the complete schema update process to prevent future issues.
Contents:
- Overview of schema caching and validation
- Automated protection via pre-commit hooks
- Step-by-step update procedures (automatic and manual)
- Troubleshooting guide for common issues
- Schema validation checks explanation
- Best practices and workflow diagrams
- Reference to AdCP PR prebid#123 format_ids changes
Key sections:
1. When to update schemas
2. How to update (automatic vs manual)
3. Schema validation checks (format_ids, critical schemas, full sync)
4. Troubleshooting common issues
5. Update workflow diagram
6. Best practices
This prevents the issue we encountered where outdated cached schemas
caused the generator to produce incorrect types.
* Sync all schemas and regenerate with latest AdCP spec
Systematic fix to ensure all cached schemas match live registry.
Changes:
1. Updated cached schemas (tests/e2e/schemas/v1/):
- get-products-request.json
- list-creative-formats-request.json
2. Regenerated all Pydantic models (src/core/schemas_generated/):
- 77 schema files updated
- All models now use latest schema definitions
- format_ids correctly uses list[FormatId] throughout
Process:
- Ran: uv run python scripts/check_schema_sync.py --update
- Verified: All schemas now in sync with adcontextprotocol.org
- Regenerated: uv run python scripts/generate_schemas.py
Validation:
✅ All schemas in sync
✅ format_ids structure validated
✅ Schema generator working correctly
✅ No errors in schema sync check
This ensures the codebase is fully synchronized with the official
AdCP spec and prevents future schema drift issues.
* Update documentation: clarify automated schema validation
Enhanced schema-updates.md to clearly explain automated protection.
Changes:
- Added 'Pre-commit Hook (Runs Automatically)' section
- Specified that validation runs on EVERY commit
- Listed exactly what gets validated (format IDs, 7 critical schemas, 30+ endpoints)
- Clarified that commits are BLOCKED if schemas out of sync
- Added clear error message example with fix instructions
This makes it crystal clear that the automation is:
1. Always running (every commit)
2. Comprehensive (all schemas checked)
3. Enforced (commits blocked on failure)
4. Actionable (clear fix instructions)
Result: Developers understand the protection is automatic and systematic.
* Fix schema adapter to handle format_ids conversion
- Handle both dict and Pydantic model cases for filters in GetProductsRequest.to_generated()
- Update test expectations to reflect AdCP PR prebid#123 (format_ids as FormatId objects)
- Tests now pass: schema adapters properly convert string format_ids to FormatId objects during roundtrip
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
---------
Co-authored-by: Claude <noreply@anthropic.com>
bokelley
added a commit
that referenced
this pull request
Mar 18, 2026
When a working group leader adds a committee document, post a notification to the WG's Slack channel so members can track updates. This was missing — only published posts triggered channel notifications. Includes isSlackConfigured() guard, APP_URL support, and mrkdwn sanitization for title in link syntax. Resolves Addie escalation #123 (Pia/Creative WG). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
4 tasks
bokelley
added a commit
that referenced
this pull request
Mar 18, 2026
When a working group leader adds a committee document, post a notification to the WG's Slack channel so members can track updates. This was missing — only published posts triggered channel notifications. Includes isSlackConfigured() guard, APP_URL support, and mrkdwn sanitization for title in link syntax. Resolves Addie escalation #123 (Pia/Creative WG). Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
bokelley
added a commit
to bokelley/salesagent
that referenced
this pull request
May 5, 2026
* Fix format_ids handling: separate request and response fields
**Problem:**
The Package model had a validator that automatically converted format_ids
(array of strings) to format_ids_to_provide (array of FormatId objects)
during model initialization. This caused validation errors when clients
sent v1.8.0 requests with format_ids as strings, because the validator
ran too early and converted the data before the server could process it.
**Root Cause:**
- AdCP v1.8.0 spec defines TWO different fields:
- `format_ids` (request): array of strings in package-request.json
- `format_ids_to_provide` (response): array of FormatId objects in core/package.json
- The Package model was trying to serve both purposes with automatic migration
- The @model_validator ran during request deserialization, breaking client compatibility
**Solution:**
1. **schemas.py**: Removed the automatic migration validator
- Added explicit `format_ids: list[str]` field for requests
- Kept `format_ids_to_provide: list[FormatId]` field for responses
- Package model now accepts both formats without auto-conversion
2. **main.py**: Added explicit conversion in create_media_buy response building
- Convert format_ids → format_ids_to_provide when building response
- Happens after request validation, before response serialization
- Removes format_ids from response (only format_ids_to_provide in output)
**Testing:**
- All existing package-related unit tests pass (20 tests)
- Manual testing confirms Package model accepts format_ids as strings
- No breaking changes to existing code paths
**Result:**
✅ Clients can now send format_ids: ["display_300x250"] (v1.8.0 format)
✅ Server accepts request without validation errors
✅ Server responds with format_ids_to_provide: [{"agent_url": "...", "id": "..."}]
✅ Full AdCP v1.8.0 compatibility maintained
* Update schemas to use FormatId objects per AdCP PR #123
- Downloaded latest package-request.json schema (now uses FormatId objects)
- Regenerated all Pydantic schemas from latest AdCP spec
- Fixed schema generator bug: format_ids now correctly uses list[FormatId] not list[str]
- Updated server comment to reflect new spec (both request and response use FormatId)
- Maintains backward compatibility with string format IDs from older clients
Per adcontextprotocol/adcp#123:
- format_ids changed from array of strings to array of FormatId objects
- FormatId structure: {agent_url: string, id: string}
- Avoids parsing ambiguity and handles namespace collisions
* Fix schema generator: regenerate with correct package-request schema
Root cause: Outdated cached schema file had format_ids as list[str] instead of
list[FormatId] with $ref to format-id.json.
Investigation findings:
1. Schema generator's resolve_refs_in_schema() function works correctly
2. Issue was outdated cached file: tests/e2e/schemas/v1/_schemas_v1_media-buy_package-request_json.json
3. Old cached schema had: items: {type: 'string'}
4. New cached schema has: items: {$ref: '/schemas/v1/core/format-id.json'}
5. After downloading latest schema, generator correctly resolved $ref to FormatId
Result:
✅ Generated schemas now correctly use list[FormatId] (lines 378-384, 712-718)
✅ FormatId structure properly inlined (lines 13-29) with agent_url and id fields
✅ Schema generator working as designed - just needed updated input files
Lesson: Always ensure cached schemas are up-to-date before regenerating.
* Enhance schema sync checker to validate format_ids structure
Adds automated validation to prevent outdated schemas from causing
schema generator issues in the future.
Changes:
1. Added package-request.json and format-id.json to critical schemas list
2. Added new validate_format_ids_structure() check that specifically validates:
- format_ids field uses $ref to /schemas/v1/core/format-id.json
- Detects if format_ids incorrectly uses {type: 'string'}
- Provides clear error messages with fix instructions
3. Runs as first check in pre-commit hook for early detection
Prevention:
This prevents the root cause issue from PR investigation:
- Old cached schema had format_ids: {items: {type: 'string'}}
- New spec requires format_ids: {items: {$ref: '/schemas/v1/core/format-id.json'}}
- Schema generator was working correctly, but had wrong input
Testing:
✅ Runs successfully: uv run python scripts/check_schema_sync.py
✅ Validates format_ids structure correctly
✅ Provides actionable error messages
Result:
Future schema updates will be caught by pre-commit hook before
causing schema generator issues.
* Add comprehensive schema update documentation
Documents the complete schema update process to prevent future issues.
Contents:
- Overview of schema caching and validation
- Automated protection via pre-commit hooks
- Step-by-step update procedures (automatic and manual)
- Troubleshooting guide for common issues
- Schema validation checks explanation
- Best practices and workflow diagrams
- Reference to AdCP PR #123 format_ids changes
Key sections:
1. When to update schemas
2. How to update (automatic vs manual)
3. Schema validation checks (format_ids, critical schemas, full sync)
4. Troubleshooting common issues
5. Update workflow diagram
6. Best practices
This prevents the issue we encountered where outdated cached schemas
caused the generator to produce incorrect types.
* Sync all schemas and regenerate with latest AdCP spec
Systematic fix to ensure all cached schemas match live registry.
Changes:
1. Updated cached schemas (tests/e2e/schemas/v1/):
- get-products-request.json
- list-creative-formats-request.json
2. Regenerated all Pydantic models (src/core/schemas_generated/):
- 77 schema files updated
- All models now use latest schema definitions
- format_ids correctly uses list[FormatId] throughout
Process:
- Ran: uv run python scripts/check_schema_sync.py --update
- Verified: All schemas now in sync with adcontextprotocol.org
- Regenerated: uv run python scripts/generate_schemas.py
Validation:
✅ All schemas in sync
✅ format_ids structure validated
✅ Schema generator working correctly
✅ No errors in schema sync check
This ensures the codebase is fully synchronized with the official
AdCP spec and prevents future schema drift issues.
* Update documentation: clarify automated schema validation
Enhanced schema-updates.md to clearly explain automated protection.
Changes:
- Added 'Pre-commit Hook (Runs Automatically)' section
- Specified that validation runs on EVERY commit
- Listed exactly what gets validated (format IDs, 7 critical schemas, 30+ endpoints)
- Clarified that commits are BLOCKED if schemas out of sync
- Added clear error message example with fix instructions
This makes it crystal clear that the automation is:
1. Always running (every commit)
2. Comprehensive (all schemas checked)
3. Enforced (commits blocked on failure)
4. Actionable (clear fix instructions)
Result: Developers understand the protection is automatic and systematic.
* Fix schema adapter to handle format_ids conversion
- Handle both dict and Pydantic model cases for filters in GetProductsRequest.to_generated()
- Update test expectations to reflect AdCP PR #123 (format_ids as FormatId objects)
- Tests now pass: schema adapters properly convert string format_ids to FormatId objects during roundtrip
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
---------
Co-authored-by: Claude <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Fixed critical schema bug where
format_idswas incorrectly defined as arrays of strings instead of structured format ID objects withagent_urlandidfields.Changes
Schema Updates (4 files)
static/schemas/v1/media-buy/package-request.jsonstatic/schemas/v1/media-buy/list-creative-formats-request.jsonstatic/schemas/v1/creative/list-creative-formats-request.jsonstatic/schemas/v1/media-buy/get-products-request.jsonAll now use
$refto/schemas/v1/core/format-id.jsoninstead oftype: "string".Documentation Updates (16 examples across 4 files)
docs/media-buy/task-reference/create_media_buy.md(10 examples)docs/media-buy/product-discovery/media-products.md(3 examples)docs/media-buy/advanced-topics/pricing-models.md(2 examples)docs/media-buy/task-reference/list_creative_formats.md(1 example)All examples now use structured format IDs:
CLAUDE.md Cleanup
Removed outdated
formats_to_providereference in favor of the consistentformat_idsfield name.Rationale
This aligns with the protocol's core requirement that format IDs are ALWAYS structured objects to:
Test Plan
🤖 Generated with Claude Code