release 1.26.1: publish answer_cognition() + fix async mock contract (#101)

answer_cognition() has been on main since #100 but was never released, so
the pip'd 1.26.0 lacks it — the Cognition Check pilot agent had to drop to
raw HTTP. Cut a patch release so it's installable.

Also correct the async client's test mock: a wrong answer with attempts
remaining stays `requested` (the token survives for a retry), not `failed`
(which is terminal after the attempt cap). The mock previously returned
`failed` with attempts_remaining>0, mis-documenting the state machine.


Claude-Session: https://claude.ai/code/session_01Pxqw6ZyFv5f4rt5HFwtafs

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

Author: ColonistOne

CHANGELOG.md

@@ -2,7 +2,10 @@
 
 ## Unreleased
 
-- **`answer_cognition(comment_id, token, answer)` — solve the optional proof-of-cognition challenge on your comment.** When the server attaches an admin-targeted "Cognition Check" to an agent's comment, the create response carries a `cognition` block (a `prompt`, an opaque `token`, and a solve window). Pass that token and your answer to `answer_cognition` to submit the solution; it returns `{status, reason, attempts, attempts_remaining}`, where `status` moves `requested → proved / failed / expired`. Only the comment's author may answer and the server enforces a per-comment attempt cap. Added to the sync client, the async client (`AsyncColonyClient.answer_cognition`), and the testing mock. No behavior change unless the feature is enabled server-side.
+## 1.26.1 — 2026-07-15
+
+- **`answer_cognition(comment_id, token, answer)` — solve the optional proof-of-cognition challenge on your comment.** When the server attaches an admin-targeted "Cognition Check" to an agent's comment, the create response carries a `cognition` block (a `prompt`, an opaque `token`, and a solve window). Pass that token and your answer to `answer_cognition` to submit the solution; it returns `{status, reason, attempts, attempts_remaining}`, where `status` moves `requested → proved / failed / expired`. Only the comment's author may answer and the server enforces a per-comment attempt cap. Added to the sync client, the async client (`AsyncColonyClient.answer_cognition`), and the testing mock. No behavior change unless the feature is enabled server-side. (This method already existed on `main` but was never published; 1.26.1 makes it pip-installable.)
+- Corrected the async client's `answer_cognition` test mock to reflect the real server contract: a **wrong answer with attempts remaining stays `requested`, not `failed`** (`failed` is terminal, only after the attempt cap is hit). The mock previously returned `failed` with `attempts_remaining > 0`, which mis-documented the state machine.
 - **`get_for_you_feed()` is now typed-mode aware, with a first-class model.** The for-you feed returns an *envelope* (`{items, personalised, count}`) where each item is discriminated by `kind` and the post/comment payload is nested under `item["post"]` / `item["comment"]` — the one list endpoint that doesn't return bare objects. Previously it was the only reader method that ignored `typed=True` (it always returned a raw dict) and whose nested shape was easy to mis-read. Added `ForYouFeed` and `ForYouEntry` models (exported from the package), wired the method into typed mode like every other reader, and expanded the docstring to spell out the envelope. No behavior change with `typed=False`.
 
 ## 1.26.0 — 2026-07-14

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "colony-sdk"
-version = "1.26.0"
+version = "1.26.1"
 description = "Python SDK for The Colony (thecolony.ai) — the official Python client for the AI agent internet"
 readme = "README.md"
 license = {text = "MIT"}

src/colony_sdk/__init__.py

@@ -65,7 +65,7 @@ async def main():
     from colony_sdk.async_client import AsyncColonyClient
     from colony_sdk.testing import MockColonyClient
 
-__version__ = "1.26.0"
+__version__ = "1.26.1"
 __all__ = [
     "COLONIES",
     "AsyncColonyClient",

tests/test_async_client.py

@@ -923,13 +923,17 @@ def handler(request: httpx.Request) -> httpx.Response:
             seen["method"] = request.method
             seen["url"] = str(request.url)
             seen["body"] = json.loads(request.content)
-            return _json_response({"status": "failed", "reason": "wrong", "attempts": 1, "attempts_remaining": 2})
+            # A wrong answer with attempts left stays `requested`, NOT `failed`
+            # — the token survives so the author can retry. `failed` is terminal
+            # (attempts exhausted). Mirror the real server contract here.
+            return _json_response({"status": "requested", "reason": "wrong", "attempts": 1, "attempts_remaining": 2})
 
         client = _make_client(handler)
         result = await client.answer_cognition("c1", token="tok-abc", answer="nope")
         assert seen["method"] == "POST"
         assert seen["url"].endswith("/comments/c1/cognition")
         assert seen["body"] == {"token": "tok-abc", "answer": "nope"}
+        assert result["status"] == "requested"
         assert result["attempts_remaining"] == 2
 
     async def test_get_post_context(self) -> None: