← all publishers

martintmk

@martintmk source repo

Report this profile Claim this profile

• 18 skills
• 0 followers
• 6 hours ago last updated

1. ▌
   Pr Review Queue · martintmk bundle

   Review eligible GitHub/ADO PRs sequentially at a confirmed cadence. Watch target-authored PRs until merge and other PRs only through their first seven days. Use for "monitor and review PRs", "review my PR queue" or recurring reviews, not discovery digests or feedback/fix loops. Installing or editing this skill does not start monitoring.
   0 installs
2. ▌
   Pr Review Radar · martintmk bundle

   Find unreviewed open PRs in specified GitHub and Azure DevOps repositories: recent PRs, foundational APIs, infrastructure, or mentions of the user. Send only previously unreported PRs to Teams self-chat. Use for review opportunities, repository monitoring, or recurring review digests; not code review, unanswered-feedback discovery, or acting on feedback.
   0 installs
3. ▌
   Review Delivery · martintmk bundle

   Deliver finished, merged review findings once to GitHub, Azure DevOps or a local report, with precise anchors and the shared AI-attributed comment format. Use when a review is ready to deliver, including a clean review, or to recover failed posting. Not for generating findings, formatting an area's intermediate output, or non-review PR comments.
   0 installs
4. ▌
   Review Telemetry · martintmk

   Review Rust changes for metrics, logs and spans that break OpenTelemetry semantic conventions, destabilize an existing dimension set, risk unbounded cardinality, allocate per emission, or duplicate instrumentation a library already provides. Use for a focused telemetry audit, or when review-lens routes changed telemetry here. Not for general logging style or for telemetry backend and exporter configuration review.
   0 installs
5. ▌
   Feedback Autonomy · martintmk

   Gate actions on feedback for the single PR an unattended session owns. Handle proven automation and instructions from the same human who created the PR, subject to impact and task authority. Finish independent actionable comments and build problems before batching remaining approval requests. Use before acting on feedback and after authorization. Not for radar discovery or generating review findings.
   0 installs
6. ▌
   Pr Feedback Radar · martintmk

   Find new unanswered human requests directed at the user on recently opened, still-open GitHub and Azure DevOps PRs. Prioritize demonstrably blocking responses and send a deduplicated Teams self-chat digest. Use for unanswered feedback, response reminders, or recurring feedback digests; not code review, acting on feedback, bot activity, or already-answered comments.
   0 installs
7. ▌
   Review API Design · martintmk

   Review a Rust change's public contract from the diff and source: visibility and layering, semver cascade, constructors, builders and defaults, strong types and enums, trait implementability and sealing, macros as public API, and error type, message and panic conventions, including internal errors. Also owns dependency and feature checks. Use for downstream contracts, manifest or error-policy changes, or review-lens's public-surface pass. Complements review-public-api, which audits an existing surface from cargo public-api output without reading source. Not for runtime defects or naming style.
   0 installs
8. ▌
   Review Public API · martintmk bundle

   Audit a Rust library's exported contract using cargo-public-api output only, then isolated rustdoc-based filtering of provisional claims. Use for a whole-crate or explicit output-only API audit, or review-lens's mandatory output-only pass; small PRs cover changed public items and their immediate family. Applies idiomatic Rust API practices and API-visible Pragmatic Rust Guidelines. Not for source-based findings, implementation correctness, docs quality/consistency, performance or posting.
   0 installs
9. ▌
   Review Resilience · martintmk

   Review Rust changes for recovery classification and retry, timeout, circuit-breaker, hedging, fallback or chaos behavior that should use approved middleware. In recoverable/seatbelt repositories, checks Recovery against the exact recoverable version's _documentation::recipes and audits seatbelt adoption. Use for a PR, branch, commit, working-tree diff, or focused crate audit. Not for a general code review or ordinary error changes with no recovery concern.
   0 installs
10. ▌
    Review Consistency · martintmk

    Review changed code and documentation for conflicting contracts, stale claims, examples or defaults, documentation gaps and conflicting documents. Use for "check code and docs agree", "review documentation consistency", or when review-lens routes a change to documented behavior here. Not for prose styling, naming preferences, runtime defect hunting or output-only API audits.
    0 installs
11. ▌
    Review Correctness · martintmk

    Review Rust changes for behavioral defects and prove each one before raising it: control-flow and version gates, boundary and representation limits, resource and free-list models, cancellation and drop safety, time arithmetic, and round-trip or invariant claims. Uses reproductions, Miri and bounded adversarial inputs under the shared verification discipline. Use for a focused defect audit, or when review-lens routes behavioral risk here. Not for API design, naming, or formatting.
    0 installs
12. ▌
    Review Public Docs · martintmk bundle

    Retrieve scoped Rust public API docs from cargo-generated rustdoc JSON for a PR, branch, commit, working tree or explicit items. Returns a compact bundle of paths, full item/member docs, attributes, links and coverage. Use for "what do these APIs document", documentation presence, or authoritative docs needed by another reviewer. Not for design, prose/consistency judgments, implementation correctness, posting or rendered docs.rs pages.
    0 installs
13. ▌
    Teams Self Message · martintmk

    Send a text or HTML message to the current user's Microsoft Teams self-chat. Use when the user says "message myself", "send me a Teams message", "send this to my Teams", or invokes this skill with message content. The message is posted as the user to the Teams "Just me" chat, so Teams may not produce an incoming-message notification.
    0 installs
14. ▌
    Minimal Local Validation · martintmk

    Keep local Rust validation intentionally narrow when the build pipeline already performs comprehensive workspace checks. Use when asked to minimize local validation, avoid long workspace-wide checks, run cargo check only for changed crates, or rely on CI for exhaustive validation and fix only concrete pipeline failures. Not for repositories without comprehensive CI, explicit full-validation requests, or diagnosing an already failing pipeline.
    0 installs
15. ▌
    Review Lens · martintmk bundle

    Review a Rust pull request, branch, commit or working-tree diff as an autonomous AI reviewing agent applying @martintmk's library-maintainer priorities. Establishes facts once, dispatches every review sub-skill to a fresh agent context, merges findings and delivers one AI-attributed review. Use for "review this PR", "review my changes" or "review like me". For a focused area, invoke its review-* skill directly. Not for formatting-only passes, output-only API audits or specialist security reviews.
    0 installs
16. ▌
    Review Perf · martintmk

    Review Rust changes for avoidable allocations, hot-path cost, and unmockable time or randomness. Covers needless String/clone/round-trip allocation, per-call dispatch that could be decided once, algorithmic growth, contention, hard-wired clocks or entropy where injectable sources exist, and honest benchmark reporting. Use for cost or clock/randomness-injection audits, or when review-lens routes a hot path or claimed optimization here. Not for micro-style preferences or off-path optimizations that add complexity for no measured gain.
    0 installs
17. ▌
    Review Tests · martintmk

    Review Rust changes for deleted or weakened tests and observable behavior deltas, even when test files are untouched. Protects existing behavioral contracts from being rewritten to match an implementation, especially in AI-authored changes, and checks that changed tests are idiomatic and reuse supported test utility features such as test-util. Use when asked to review test preservation, behavior compatibility, or Rust test quality in a PR, branch, commit, working-tree diff, or focused audit. Not for coverage percentages or test-framework formatting.
    0 installs
18. ▌
    Review Naming · martintmk

    Review Rust changes for names and shapes that diverge from their siblings, and for abstractions that earn nothing. Covers family and convention alignment, concise non-padded names, units already carried by a type, domain-appropriate terminology, and whether a new trait, wrapper or layer could be replaced by a small change to an existing type. Use for a focused naming or abstraction audit, or when review-lens routes new names here. Not for formatting, import order, or public contract and semver review.
    0 installs