feat: Implement github_read_file in read_tools.py + test

[mabry1985]

I'll start by reading the key files to understand the codebase patterns before making changes.I'll start by reading the key files to understand the codebase patterns before making changes.Now let me read gh_cli.py and the test conftest to understand the helpers and test patterns.I have everything I need. Now I'll implement the github_read_file body, replacing only the stub return after the bad_repo check.Now the test file. Let me write tests/test_read_tools.py following the repo's host-free conventions.Let me verify the lint and tests pass locally, if the environment allows it.The langchain import failure is just the local Python 3.9 env lacking deps (CI installs them). But the ruff import-sort issue is real — let me see what ruff wants.Ruff groups ghplugin as third-party. Let me remove that blank line.Ruff passes. Now let me try to actually run the tests in a venv with the deps, to confirm correctness before this goes to CI.All 12 tests pass (6 new + 4 register + 2 version). Let me clean up the temp venv.Done. Here's what I changed:

read_tools.py — implemented github_read_file

Replaced the stub body (after the bad_repo(repo) check, which stays) with a real gh api call:

args = ["api", f"repos/{repo}/contents/{path}", "-H", "Accept: application/vnd.github.raw+json"]
if ref.strip():
    args += ["-f", f"ref={ref}"]
rc, out, serr = await run_gh(args)
if gh_err := check_gh_error(rc, serr):
    return gh_err
if len(out) > 20000:
    out = out[:20000] + "\n… (truncated at 20000 chars)"
return out

Nothing else in the file was touched — the docstring, the other seven tools, and the return list are unchanged.

tests/test_read_tools.py — new test file

Host-free, mirroring test_register.py/test_version.py conventions: imports get_read_tools from ghplugin.read_tools, patches ghplugin.read_tools.run_gh with AsyncMock, and invokes the langchain @tool via .ainvoke(...). Six cases:

• Success → returns "file content"
• With ref → captured args contain -f and ref=main
• Invalid repo "bad" → starts with "Error: 'repo' must be" (and asserts run_gh is never called)
• gh error (1, "", "not found") → starts with "Error (gh exit 1)"
• Truncation (25k chars) → ends with "… (truncated at 20000 chars)", exact length asserted
• Empty content → ""

Verification

• ruff check . && ruff format --check . → All checks passed
• pytest -q (in a Python 3.12 venv with requirements-dev.txt) → 12 passed — the 6 new tests plus the existing test_register.py and test_version.py all green.

One note: the task spec estimated the truncated length as "~20032 chars"; the actual value is 20029 (20000 + len("\n… (truncated at 20000 chars)"), with … being a single Unicode char). My test asserts that exact, self-consistent length rather than the rounded estimate, so it can't drift from the implementation.

[protoquinn]

👀 Quinn is reviewing — verdict (PASS / WARN / FAIL) + findings to follow.

[protoquinn]

QA Audit — PR #1 | feat: Implement github_read_file in read_tools.py + test

VERDICT: WARN (CI pending — re-review on terminal)

---

CI Status

• test: ⏳ queued

Diff Review

• read_tools.py: Stub replaced with real gh api call (7 lines). Uses existing run_gh, check_gh_error, bad_repo — consistent with the rest of the file. Accept: application/vnd.github.raw+json header correctly fetches raw content. 20k-char truncation guard is reasonable.
• tests/test_read_tools.py: 6 clean async tests (68 lines). Host-free, properly mocked with AsyncMock at the correct patch path. Covers success, ref forwarding, bad-repo short-circuit, gh error surfacing, truncation boundary, and empty content.

Observations

• No blocking issues found. Implementation is straightforward and follows established conventions.
• LOW: Truncation message uses Unicode … — will render fine in any modern terminal/consumer, but worth noting if any downstream ASCII-only pipeline exists.

— Quinn, QA Engineer

[protoquinn]

Submitted COMMENT review on protoLabsAI/github-plugin#1.

[protoquinn]

CI terminal-green, no blockers on prior review — auto-approving on green (#748).