coding-standards

Plugin coding-standards from marketplace coding-standards in ripta/coding_agent_standards. Commands read from the repo's manifest — 3 days ago.

Skills (34)

• agent-router

  Use this skill when the user needs help choosing which agent or skill to use, or wants to understand what's available.

• architecture-exploration

  Use this skill to produce a deep, citation-grounded ARCHITECTURE.md for the current project by investigating the codebase. Detects the project type (backend service, frontend app, CLI, library, or full-stack) and tailors the reference to it. If invoked with arguments, scopes the exploration to that

• bug-triage

  Verify the oldest open bug in the project's bugs directory. Reproduce it against the current build, then close it if fixed. If it still reproduces, estimate the fix's size and design decisions, then recommend fixing it now or opening a proposal. Reconciles the bug index when a bug file is already cl

• cloudbuild

  Author, audit, or fix a Google Cloud Build `cloudbuild.yaml`, and create the matching GCB triggers. Use when a repo needs CI on Cloud Build, when an existing cloudbuild.yaml has drifted from standard, or when a repo has a config but no triggers.

• conversation-handoff

  Use this skill to compact the current conversation into a standalone markdown handoff document that another coding agent / LLM can use to resume the work. Detects external references (proposals, specs, source files, ADRs) and asks whether to inline them (comprehensive) or cite them by name (compact)

• data-visualization

  Use this skill before making any chart, graph, or plot, in any medium or library: an HTML page, a slide, a notebook, a PNG for a report, or a terminal plot. It sets what a chart claims and whether a reader can trust it: titles, axis labels and units, zero baselines, truncated and zoomed ranges, cons

• debug-ci

  Diagnose CI failures on Google Cloud Build for a PR. Automatically finds the PR, downloads build logs, identifies the failure category, and either suggests a fix (for timeouts) or enters plan mode to debug real failures.

• dependabot-sweep

  Sweep a GitHub owner's open dependabot PRs across all its repositories. Verifies each PR really is dependabot's, sorts them into buckets, and drives each one to merged or to a reported dead end. Triggers: "scan my repos for PRs awaiting review", "sift through dependabot PRs", "merge the dependabot b

• format-checker

  Use this skill to check or fix code formatting. Includes pre-commit checks, CI formatting failures, and selective formatting of specific languages.

• gitignore-maintainer

  Use this skill to add, remove, or review .gitignore patterns.

• go-code-reviewer

  Use this skill after writing or modifying Go code to get a thorough review for best practices, security, stability, and consistency with the codebase.

• init-zaz

  Scan the current project and generate a zaz.toml config file. Use when bootstrapping a new project to run under zaz. Reads the zaz configuration reference and the worked examples bundled with this skill.

• land-branch

  Land a branch, or the branch up to a given commit, into the mainline through PRs, without touching the user's branch. Snapshots it onto an agent-owned staging branch in a worktree, splits a large range into batches that land concurrently when they are independent, and runs the merge-fix loop on each

• lint-runner

  Use this skill to run linters, check code quality, or diagnose linting failures. Covers all languages and linting tools configured in the project.

• makefile-helper

  Use this skill to find the right Makefile command, explain what a target does, list available targets, or audit a Makefile against the Makefile standards. Helps users navigate the project's build system. Audits are report-only.

• merge-fix-loop

  Drive a branch to merged. Pushes an agent-owned branch, opens a PR, enables auto-merge, then loops: waits on CI, triages and fixes failures, triages and answers review comments, until the PR merges or the time budget runs out. Triggers: "merge-fix loop", "get this PR merged", "babysit this PR", "pus

• new-proposal

  Create a new proposal from a short description, following the project's proposal standards. Determines the proposals directory, project prefix, and next sequential number, interviews the user to flesh out the design, and writes a draft proposal. Invoked as `/new-proposal <description>`; the descript

• open-questions-audit

  Scan a project's in-flight proposals (those not yet finalized or closed) and rank them by the number of unresolved design questions. Reports the top proposals with the most open questions so design review can be prioritized.

• presentation

  Use this skill to build a terminal slide deck for presenterm from existing source material: research, proposals, ADRs, ARCHITECTURE.md, design docs, implementation reports, or code. It inventories how much detail the sources hold, asks how much the main deck and the appendix should carry, then write

• proposal-cleanup

  Reconcile a project's proposals, proposal index, and queue after milestones finish. Marks proposals implemented when every milestone is finished, syncs the index, moves finished queue rows to history.md, and reports drift it may not fix. Also trims the proposal dependency graph and prunes captured i

• proposal-reviewer

  Review a proposal's open design questions, research options, discuss tradeoffs, and record decisions. Completes the proposal pipeline: research -> draft -> design review -> accepted. Runs manual, converge-then-ask, or fully autonomous with critic convergence checks.

• rca

  Root cause analysis from a set of symptoms. Builds hypotheses, tests each one against evidence, narrows to the likely causes, then asks "why" down each cause until the chain stops. Every link carries its evidence and a marker for its confidence. Triggers: "root cause", "rca", "why did this happen",

• release

  Use this skill to cut a tagged git release. Auto-detects the project's tagging scheme—month-based (vYYYY.M.N, where N is the Nth tag that month) or semver (vX.Y.Z)—proposes the next tag (recommending a patch/minor/major bump for semver), drafts release notes from the commit delta, and creates an ann

• research-audit

  Validate existing research, often someone else's, given as local markdown, PDF, or HTML files, URLs, or pasted text. Checks each factual claim against its sources, rebuilds the problem from verified facts, researches the gaps, and optionally challenges the solutions the research prescribes. Triggers

• research-to-proposals

  Review project research documents and generate proposals based on findings. Point it at a directory of research docs and it will read them, identify actionable items, and create draft proposals following the project's proposal standards.

• risk-assessment

  Use this skill to assess the risk of a code change before merging or deploying. Scores five risk dimensions (rollback risk, blast radius, regression history, silent-failure risk, novelty) as Low/Med/High with rationale, backed by deterministic git metrics from a companion script. Targets the working

• schedule-proposal

  Schedule accepted proposals (PROJ-NNN) by adding their milestones to the project's queue. With no ID, recommends which to schedule and asks. Spec-only: edits queue.md and the proposals' Decision Logs. No code planning or implementation. Invoked as `/schedule-proposal [PROJ-NNN ...]`.

• session-report

  Write the current session's findings into a durable, self-contained, reproducible report file in the user's reports directory. Triggers: "write this up as a report", "write a report of these findings", "save/write this to my reports directory", "write the session findings to a file". DO NOT trigger

• test-audit

  Audit a scope of tests for value against cost. Finds low-value proofs, groups them into coherent batches that keep one owner per contract, asks which batches to apply, applies them, and reports cost before and after. Invoked as `/test-audit <path or package>`.

• testing-specialist

  Use this skill when writing tests, choosing between a unit and an integration test, deciding whether a test is worth its cost, or debugging test failures and flaky tests.

• tool-wrapper-maintainer

  Use this skill to update tool versions, add new tools, or maintain versioned wrapper scripts in the bin/ directory.

• tracker-sync

  Reconcile issue-tracker groups (Jira epics, Linear projects, GitHub tracking issues, milestones, or labels) with work merged in mapped repository subdirectories since the last sync. Closes or updates tickets whose work landed, proposes tickets for untracked proposals or queue rows, fixes proposal an

• tune-reviewer

  Turn a style or review correction into durable agent instructions. Give it a git range, a commit, a diff, or a plain-English description of a correction the code reviewer should have caught. It names the underlying principle, checks what is already covered, and adds the genuinely-new part to the rev

• walkthrough

  Use this skill to produce a self-contained interactive HTML page that explains a concept, feature, changeset, commit, diff, branch, or PR. It gathers evidence from the repo, verifies every claim with fast subagents, then writes one static HTML file with diagrams, annotated code, interactive figures,

Slash commands (9)

• /coding-standards:converge-design

  Iterate a design question with the design critic until it converges, then ask

• /coding-standards:guardian

  Own the changes in a domain, and handle change requests from other sessions

• /coding-standards:migrate-era

  Migrate a project's proposals and phases to era 3, after you approve the plan

• /coding-standards:next-pr

  Find the next chunk of unmerged commits to merge into the mainline branch

• /coding-standards:pr-description

  Write a PR title and description for the changes on this branch

• /coding-standards:summary

  Summarize this conversation's work as a commit message

• /coding-standards:update-profile

  Add or replace a project's coding-standards profile import, and fix imports of the old checkout paths

• /coding-standards:work-on

  Plan and implement a proposal milestone

• /coding-standards:work-queue

  Plan a run of queue rows up front, then implement and commit their milestones one at a time

Subagents (7)

• adr-impact

  Checks a change or a proposal against the project's ADRs. Reports each ADR it touches as consistent, superseding, or in conflict, with the ADR text cited. Read-only. Runs beside the reviewer.

• citation-verifier

  Checks a batch of file:line citations against the source, in the working tree or at a git revision. Returns one verdict per claim: CONFIRMED, CORRECTED with the right line, REFUTED, or UNCHECKABLE. Read-only. It does not fix, improve, or editorialize on the claims.

• claim-verifier

  Checks a batch of factual claims from a research document against their sources: web pages, documents, or a search when a claim cites nothing. Returns one verdict per claim: CONFIRMED, CORRECTED, REFUTED, OUTDATED, UNSOURCED, or UNCHECKABLE. Read-only. It does not fix, improve, or editorialize on th

• design-critic

  Critiques a design before it is built. Ranks the options for a design question on its own, finds missing options, and tries to break the recommended one. Also critiques a whole proposal before it is accepted. Read-only.

• design-judge

  Breaks a tie on a design question after repeated critique failed to converge. Reads the finalists and the argument history from every pass, and rules on one option, or returns the question to the user when it is truly theirs to decide. Read-only.

• implementer

  Implements one proposal milestone in the working tree from a coordinator's brief. Runs the local checks and reports what it changed. It does not commit.

• reviewer

  Reviews the working-tree diff after a proposal milestone, or a commit range spanning several milestones. Covers plan-conformance, proposal amendments and the freeze, correctness/security, and comment craft, plus ADRs on a final review. Reports findings; it does not edit code.

Hooks (14)

• SessionStart

  "${CLAUDE_PLUGIN_ROOT}"/hooks/link-current.sh

• SessionStart

  "${CLAUDE_PLUGIN_ROOT}"/hooks/revisit-notice.sh

• SessionStart

  "${CLAUDE_PLUGIN_ROOT}"/hooks/check-setup.sh

• SessionStart

  "${CLAUDE_PLUGIN_ROOT}"/hooks/guardian-session-start.sh

• SessionStart

  "${CLAUDE_PLUGIN_ROOT}"/hooks/standards-on-demand.py

• SessionEnd

  "${CLAUDE_PLUGIN_ROOT}"/hooks/guardian-session-end.sh

• UserPromptSubmit

  "${CLAUDE_PLUGIN_ROOT}"/hooks/name-guardian.sh

• PreToolUse

  "${CLAUDE_PLUGIN_ROOT}"/hooks/block-broad-find.sh

• PreToolUse

  "${CLAUDE_PLUGIN_ROOT}"/hooks/block-rm-rf.sh

• PreToolUse

  "${CLAUDE_PLUGIN_ROOT}"/hooks/block-private-terms.sh

• PreToolUse

  "${CLAUDE_PLUGIN_ROOT}"/hooks/block-gcloud-project-number.sh

• PreToolUse

  "${CLAUDE_PLUGIN_ROOT}"/hooks/standards-on-demand.py

• PostToolUse

  "${CLAUDE_PLUGIN_ROOT}"/hooks/format-edited-file.sh

• Stop

  targets=$(for t in test lint; do make -n "$t" >/dev/null 2>&1 && printf " %s" "$t"; done); if [ -n "$targets" ]; then echo "Reminder: run make$targets before considering work complete"; fi

Similar plugins