feat: add with_metadata option to Prompts.get()

[carlos-marchal-ph]

💡 Motivation and Context

Port of PostHog/posthog-js#3387 to Python.

Prompts.get() returns only the prompt string, discarding name and version from the API response. Users want to track $ai_prompt_name / $ai_prompt_version on AI generation events without a separate HTTP call.

Adds a with_metadata option to get(). When True, returns a PromptResult dataclass with source (api, cache, stale_cache, or code_fallback), name, and version. Deprecates calling get() without with_metadata (warns once per instance via DeprecationWarning).

💚 How did you test it?

• 14 new unit tests covering all metadata sources, deprecation warnings, cache sharing, and strengthened API response validation
• All 51 tests pass (21 existing unchanged + 14 new + 16 compile/cache)

📝 Checklist

• I reviewed the submitted code.
• I added tests to verify the changes.
• I updated the docs if needed.
• No breaking change or entry added to the changelog.

If releasing new changes

• Ran sampo add to generate a changeset file
• Added the release label to the PR

[github-actions]

posthog-python Compliance Report

Date: 2026-04-15 15:33:16 UTC
Duration: 159995ms

⚠️ Some Tests Failed

29/30 tests passed, 1 failed

---

Capture Tests

✅ 29/29 tests passed

View Details

Test	Status	Duration
Format Validation.Event Has Required Fields	✅	517ms
Format Validation.Event Has Uuid	✅	1508ms
Format Validation.Event Has Lib Properties	✅	1507ms
Format Validation.Distinct Id Is String	✅	1507ms
Format Validation.Token Is Present	✅	1509ms
Format Validation.Custom Properties Preserved	✅	1507ms
Format Validation.Event Has Timestamp	✅	1507ms
Retry Behavior.Retries On 503	✅	9515ms
Retry Behavior.Does Not Retry On 400	✅	3511ms
Retry Behavior.Does Not Retry On 401	✅	3507ms
Retry Behavior.Respects Retry After Header	✅	9515ms
Retry Behavior.Implements Backoff	✅	23530ms
Retry Behavior.Retries On 500	✅	7501ms
Retry Behavior.Retries On 502	✅	7517ms
Retry Behavior.Retries On 504	✅	7508ms
Retry Behavior.Max Retries Respected	✅	23535ms
Deduplication.Generates Unique Uuids	✅	1497ms
Deduplication.Preserves Uuid On Retry	✅	7512ms
Deduplication.Preserves Uuid And Timestamp On Retry	✅	14519ms
Deduplication.Preserves Uuid And Timestamp On Batch Retry	✅	7511ms
Deduplication.No Duplicate Events In Batch	✅	1508ms
Deduplication.Different Events Have Different Uuids	✅	1508ms
Compression.Sends Gzip When Enabled	✅	1508ms
Batch Format.Uses Proper Batch Structure	✅	1507ms
Batch Format.Flush With No Events Sends Nothing	✅	1006ms
Batch Format.Multiple Events Batched Together	✅	1506ms
Error Handling.Does Not Retry On 403	✅	3508ms
Error Handling.Does Not Retry On 413	✅	3509ms
Error Handling.Retries On 408	✅	7510ms

Feature_Flags Tests

⚠️ 0/1 tests passed, 1 failed

View Details

Test	Status	Duration
Request Payload.Request With Person Properties Device Id	❌	505ms

Failures

request_payload.request_with_person_properties_device_id

404, message='NOT FOUND', url='http://sdk-adapter:8080/get_feature_flag'

[greptile-apps]

Comments Outside Diff (1)

1. posthog/ai/prompts.py, line 65-72 (link)

   bool passes isinstance(..., int) check

   bool is a subclass of int in Python, so isinstance(True, int) returns True. A response body with "version": true from a misconfigured server or test stub would silently pass validation and be stored as version=True in CachedPrompt and returned in PromptResult. Use type(...) is int to exclude booleans:

   Prompt To Fix With AI

   This is a comment left during a code review.
   Path: posthog/ai/prompts.py
   Line: 65-72
   
   Comment:
   **`bool` passes `isinstance(..., int)` check**
   
   `bool` is a subclass of `int` in Python, so `isinstance(True, int)` returns `True`. A response body with `"version": true` from a misconfigured server or test stub would silently pass validation and be stored as `version=True` in `CachedPrompt` and returned in `PromptResult`. Use `type(...) is int` to exclude booleans:
   
   
   
   How can I resolve this? If you propose a fix, please make it concise.

Prompt To Fix All With AI

This is a comment left during a code review.
Path: posthog/ai/prompts.py
Line: 65-72

Comment:
**`bool` passes `isinstance(..., int)` check**

`bool` is a subclass of `int` in Python, so `isinstance(True, int)` returns `True`. A response body with `"version": true` from a misconfigured server or test stub would silently pass validation and be stored as `version=True` in `CachedPrompt` and returned in `PromptResult`. Use `type(...) is int` to exclude booleans:

```suggestion
def _is_prompt_api_response(data: Any) -> bool:
    """Check if the response is a valid prompt API response."""
    return (
        isinstance(data, dict)
        and isinstance(data.get("prompt"), str)
        and isinstance(data.get("name"), str)
        and type(data.get("version")) is int
    )
```

How can I resolve this? If you propose a fix, please make it concise.

---

This is a comment left during a code review.
Path: posthog/ai/prompts.py
Line: 211-213

Comment:
**Warn-once flag is not thread-safe**

The read-then-write on `_has_warned_deprecation` is not atomic. Under concurrent access, two threads can both observe `False` before either sets it to `True`, resulting in duplicate warnings. Python's standard library uses a threading lock for this pattern (e.g. `threading.Lock` or `warnings.warn` with a filter). For a deprecation-only warning the practical impact is low, but if thread safety matters here a `threading.Lock` or calling `warnings.warn(..., stacklevel=2)` without the manual flag (relying on the warning module's own per-location deduplication) would be safer.

How can I resolve this? If you propose a fix, please make it concise.

_{Reviews (1): Last reviewed commit: "chore: add sampo changeset for prompt me..." | Re-trigger Greptile}