Python good first issues

AI Summary: The integration is incorrectly reading insulation resistance and leakage current data from the wrong Modbus registers. It's currently fetching serial number text from registers 3087-3091, which are then misinterpreted as sensor readings. The correct registers (200-205) appear to contain live, measurable data but are not currently mapped.

AI Summary: This issue proposes to implement MCP tool orchestration for the `/ask` endpoint, replacing a previous stalled effort. The implementation will be limited to streamable HTTP servers provided by the operator, with a focus on security by rejecting untrusted forks from enabling MCP features. Several components from the previous attempt have been deliberately excluded for specific reasons.

AI Summary: The project's `uvicorn` dependency is pinned to an outdated version (0.22.0) since its initial commit, preventing the use of newer versions required by the MCP SDK. This issue proposes relaxing the pin to a more recent version, with evidence showing minimal impact on tests and local builds.

AI Summary: This issue requests support for reading GRIB edition 0 (GRIB0) files embedded within Météo-France FA container files, which VirtualiZarr currently cannot handle. It outlines two main challenges: decoding GRIB0 messages and parsing the FA container format, proposing three potential implementation strategies.

AI Summary: The user is reporting a bug where the system stops responding after processing a long context. This is indicated by an image showing the issue and a Russian comment stating "I understand that after a long context it just doesn't respond." The issue is labeled as a bug and a "good first issue."

AI Summary: The GEMM-ceiling probe incorrectly hardcodes the use of bfloat16, leading to pessimistic throughput forecasts on pre-Ampere GPUs. These older cards lack native bfloat16 tensor cores, so the probe measures emulated throughput instead of the actual fp16 throughput used during training. This cosmetic issue makes the pre-flight panel's forecast unreliable for a significant generation of GPUs.

AI Summary: The WebhookOutcomeStorage currently accepts invalid timeout values like NaN, infinity, and booleans, leading to unexpected behavior at runtime. This issue requires adding strict validation to the constructor and ensuring that YAML configurations with invalid timeouts raise a clear ConfigError.

AI Summary: The system incorrectly interprets quoted string values like '"false"' as true for audit receipt flags due to Python's truthiness evaluation. This leads to unexpected behavior where receipts are enabled when they should be disabled. The fix involves validating that only actual YAML booleans are accepted for these flags, raising an error for other types.

AI Summary: The `authority_window.clock_skew_tolerance_seconds` configuration currently accepts non-finite numerical values like NaN and Infinity, which can lead to runtime errors. This issue proposes validating that the clock skew tolerance is a finite, non-negative number during configuration loading and direct policy construction, rejecting invalid inputs to prevent runtime exceptions.

AI Summary: The `use_time_currency` feature incorrectly accepts non-finite values like NaN and Infinity for `max_age_seconds`. This can lead to facts never becoming stale, as comparisons with these values behave unexpectedly. The issue proposes adding validation to reject these non-finite inputs in both YAML configurations and direct Python API calls, ensuring `max_age_seconds` is always a valid, finite number or None.

AI Summary: The configuration parser for `transition:` silently accepts and coerces invalid scalar values instead of raising an error. This can lead to unexpected behavior in critical settings like agent identity and timing, potentially impacting system safety. The issue proposes stricter validation for identity fields, boolean types, and non-finite float values, along with corresponding regression tests.

AI Summary: This issue identifies a critical code duplication problem within the SSRF host predicate. Two near-identical copies of the `_is_private_or_link_local` function and four copies of the `_LOOPBACK_HOSTS` allowlist exist, leading to past security vulnerabilities where fixes were not applied consistently across all copies. The goal is to consolidate these duplicated functions and lists to ensure consistent security.

AI Summary: This issue addresses security vulnerabilities identified by CodeRabbit in a previous pull request. It requires pinning action versions, reviewing and restricting repository-wide and individual action permissions, and updating action and tooling versions. The work is recommended to be done in a single pull request after a specific prior issue is merged.

AI Summary: The "new stories collected" popup in the Assess feature is currently appearing even when no new items are actually collected (e.g., when all items are skipped). This issue proposes to modify the popup's behavior to only display when new stories have genuinely been gathered.

AI Summary: This issue proposes extending the missingness support, initially implemented for the Kalman filter, to other Gaussian filters that assume an additive Gaussian observation. The goal is to make this functionality more broadly applicable across the library.

AI Summary: The command-line interface (CLI) and dashboard are still displaying raw p-values with excessive decimal places, leading to values like 1e-12 being shown as '0.0000'. This issue requires updating these two locations to use the same p-value formatting convention as the Explorer, which includes a floor of '< 0.001' and three decimal places otherwise.

AI Summary: The `api-checks` CI job is significantly slowed down by a single pytest step that takes around 15-18 minutes. The proposed solution is to leverage `pytest-xdist` to parallelize the tests across multiple CPU cores, which is expected to drastically reduce the execution time. This approach is considered safe due to the existing per-test isolation mechanisms.

AI Summary: This issue is a student project invitation to compare two open-source Python implementations of the Critical Line Algorithm (CLA): the original Bailey-López de Prado implementation and the more recent `cvxcla`. The goal is to independently re-measure and validate the findings presented in the existing `comparison.md` document, potentially uncovering new insights or revising current conclusions.

AI Summary: The server logs the full database connection URL, including the password, during startup. This is a security concern as the password becomes visible in container logs and any system that aggregates them. The suggested fix is to mask the password in the logged URL.

AI Summary: The README for the bluebird-api project contains a bug in its quickstart guide. It instructs users to select a scenario named 'test1' under 'Springfield', but this scenario does not exist in the provided list. This prevents users from completing the quickstart as written.

AI Summary: This issue requests documentation on how to run the project's test suite locally. It should be added to the CONTRIBUTING.md file and include instructions for running all tests from the root, as well as tests for individual packages, and any necessary setup beyond a standard development installation.

AI Summary: The `write_paths` grant in the `linux_bwrap` sandbox backend is silently ignored if the target directory does not exist on the host. This prevents agents from creating necessary output directories, leading to 'Read-only file system' errors when they attempt to write to them. The issue occurs because `bubblewrap` uses `--bind-try` for non-existent paths, which skips them without error, but the sandbox's current directory is often bound read-only.

AI Summary: This issue proposes updating the documentation for `HookContext.input_features` to accurately reflect that it now provides the exact input features resolved at planning time for the common case, rather than a best-effort runtime approximation. The change is motivated by a recent PR that improved how feature resolution is propagated.

AI Summary: This issue proposes adding a new data source for Augustus Gelatery, an Australian gelateria. The store finder is readily available and doesn't require customisation, making it a straightforward addition. The data is expected to have coordinates and opening hours.

AI Summary: This issue addresses a bug where GitLab and Gitea generate incorrect link anchors for line ranges when the start line is greater than the end line. A fix similar to the one implemented for GitHub is needed, with a suggestion for a more robust solution involving a base class helper method.

AI Summary: The Azure DevOps provider's `get_latest_commit_url` method fails with an `IndexError` when a pull request has no commits, which can occur after a force push. This issue is similar to a previously fixed bug in the GitHub provider and requires adding a check for an empty commit list before attempting to access the first element.

AI Summary: The Bitbucket Server integration incorrectly reports all repositories as having only YAML files. This is due to a hardcoded return value in `BitbucketServerProvider.get_languages` that causes all non-YAML files to be misclassified, disabling proper language-based file prioritization for Bitbucket Server pull requests. A simple fix involves returning an empty dictionary, which would allow the system to correctly categorize files as 'Other'.

AI Summary: The `/review` command incorrectly attempts to set labels even when the underlying Git provider does not support label retrieval. This leads to either exceptions being caught and ignored or empty labels being "published" without actual effect. The fix involves adding a `is_supported("get_labels")` check before label computation and publishing, similar to other label-related commands.

AI Summary: This issue proposes adding a user feedback mechanism to the deep-skill-finder project. The current implementation saves feedback to a local JSONL file, but this format causes parsing errors when multiple records are present, preventing retries. The task is to fix this so that multiple feedback records can be retried correctly.

AI Summary: This issue proposes versioning the key scheme within stored replay values. This will allow the system to distinguish between actual request divergences and changes in how keys are derived, preventing old recordings from being falsely flagged as mismatches when the derivation logic is updated.