kendex.ai92

Marketplaces / vanillagreencom/kendex

kendex

Featured

The official kendex catalog: agents, skills, hooks, commands and Pi extensions for AI coding harnesses

github.com/vanillagreencom/kendex· by vanillagreen· MIT· @ 9574cc8

Subscribe

$kendex marketplace subscribe vanillagreencom/kendex

Its packages then sit beside everything else you have. Every install is checked on your machine first.

Bundles

8

Groups of packages that work well together.

Packages

88
PackageKind
auditor

Auditor for source evidence against a stated standard, scope and period.

agent
cpa

Finance and tax preparer for work that needs the user's decision.

agent
frontend

Declarative UI specialist for TypeScript/React web, mobile and terminal views, plus all Quickshell QML and its JavaScript, view and non-view. Use for view layers, their UI messages and Quickshell QML, not non-UI shell, Python, TypeScript or Go runtime logic, or Iced.

agent
iced

Iced UI specialist. Use for Iced widgets, Canvas/Shader rendering, pane_grid layout, Theme system, Subscription-based data flow, or Elm Architecture patterns.

agent
legal

Legal preparer for a sourced answer that needs the user's decision.

agent
maintainer

Maintenance specialist for documentation, stale references, links, lint, and file or configuration organization. Use for changes settled by reading.

agent
operator

Operator for one approved browser or account step that needs the user's access.

agent
planner

Planning specialist that explores requirements and code context, weighs architecture trade-offs, and produces ordered implementation plans or plan files. May write planning artifacts; does not edit production code.

agent
publicist

Publicist for product videos, social posts and outreach drafts, with user approval recorded before each send.

agent
researcher

Exa-powered research specialist for producing evidence-backed findings reports from project research prompts. Use for research issues, technology investigations, vendor/library comparisons, architectural option analysis, and current-state web research.

agent
reviewer-arch

Architecture reviewer for design reviews, module boundary validation, spec/proposal review, and technical debt assessment.

agent
reviewer-correctness

Broad correctness and regression reviewer for behavior breakage, boundary/edge-case predicates, API/CLI/devex regressions, feature-gate leaks, migrations, state semantics, cross-module side effects, and changed UI views' behaviour, accessibility included.

agent
reviewer-doc

Documentation accuracy reviewer. Verifies changed doc claims against implementation, re-derives transcribed values, checks citations resolve, audits drift.

agent
reviewer-error

Silent failure and error handling reviewer. Detects fail-open paths, swallowed errors, wrong-cause diagnostics, and inadequate error propagation.

agent
reviewer-perf

Performance validation specialist. Latency validation, benchmark execution, percentile analysis, hot-path cost review, regression detection.

agent
reviewer-quality

Code quality reviewer for maintainability, simplification, abstraction value, type boundaries, helper reuse, decomposition, and god objects.

agent
reviewer-safety

Memory, thread, and process safety auditor. Unsafe code, data races, lock-free correctness, and file/process races (TOCTOU, PID reuse, shared mutable state).

agent
reviewer-security

Application security reviewer. Auth logic, input handling, trust/ownership gating, path containment, and secret exposure.

agent
reviewer-test

Test protection and cost reviewer. Repairs needed evidence, suggests removal of tests with no protection, and accepts justified omissions.

agent
runtime

Runtime specialist for non-UI shell, Python, TypeScript and Go code. Use for scripts, services, automation and Pi extensions.

agent
rust

Rust implementation specialist for non-Iced code. Use for domain logic, systems programming, and project-defined performance work.

agent
scout

Fast reconnaissance agent for exploring codebases, finding files by pattern, searching keywords, answering architecture questions, and returning compressed cited context or report artifacts. Specify thoroughness: quick, medium, or very thorough.

agent
swift

Swift application specialist for SwiftUI and UIKit views, Swift app code, and Xcode and Swift Package Manager builds/tests. Excludes non-UI runtime and data persistence. React Native and Expo go to frontend.

agent
tpm

Technical Program Manager for analyzing roadmaps, project lifecycle, and progress. Returns recommendations only. Does not modify project management tools.

agent
agent-runtime

Load to launch, inspect, deliver to or stop one coding-harness session through the agent-runtime CLI or Rust library.

skill
bot-instructions

Load to render, check, or adopt a repo's GitHub review-bot instruction files from the shared doctrine and its effective manifest.

skill
change-scope

Load to classify a diff's change class, or to read the shared path rules and branch measurement a tier, review depth or CI selection rests on.

skill
code-quality

Load for any coding or development task in any repository: writing, changing, fixing, refactoring, or testing code or scripts in any language.

skill
commit-guards

Load to add, tune, or debug a commit guard lane, its git hooks, or COMMIT_GUARDS_* settings.

skill
communication

Load before writing a message to a person: a reply, report, briefing, decision request, or work-item description or comment.

skill
decider

Load to create, search, or supersede an architecture decision record.

skill
deep-research

Load for research tasks, architectural investigations, and vendor, library, or technology comparisons.

skill
dep-radar

Load to run or tune a dependency sweep.

skill
dev

Load when implementing an issue or applying review fixes as a dev agent.

skill
doc-limits

Load to add, tune, or debug document byte ceilings and DOC_LIMITS_* settings.

skill
docs-writing

Load when creating or editing any Markdown file or documentation HTML, such as README, DEVELOPMENT, architecture docs, reference docs, SKILL.md and AGENTS.md, and when rewriting or reviewing one.

skill
github

Load to work or watch GitHub pull requests: threads, comments, reviews, CI logs, merges.

skill
github-repository-settings

Load to manage repository review instruction files or report live GitHub review and merge settings.

skill
harness-ci

Load to wire, tune, or debug a repo's changed-file CI skip.

skill
iced-rs

Load whenever building or debugging an Iced UI.

skill
jarvis

Load when acting for the user on their computer, browser, accounts, messages or records, or when starting, handing off or ending a long session.

skill
linear

Load for any Linear read or write: issues, projects, cycles, milestones, initiatives, labels.

skill
mailbox

Load to send, read or resolve repository mailbox messages and owner questions.

skill
orch

PRIMARY AGENT ONLY. Load to orchestrate a Linear or GitHub work item from preparation through merge.

skill
preflight

Load to run, tune, or debug preflight.

skill
price-handling

Load when comparing, rounding, formatting, or parsing prices, or designing price types.

skill
project-management

Load to plan a cycle, audit issues, build a roadmap, or decompose research into issues.

skill
review-gate

Load to adopt or operate consumer refresh.

skill
reviewer

Load when reviewing a diff, classifying findings, or returning a verdict.

skill
second-opinion

Load for a cross-model review, challenge, audit, or quick consult.

skill
slack

Load to bind an overseer's mailbox to a private Slack channel, run or check the relay, or post an alert or a file to Slack.

skill
social-media-content-calendar

Load to plan a social media campaign, build a content calendar, or adapt existing material into post drafts.

skill
worktree

Load to create, list, remove, push, or repair a git worktree.

skill
xcode-run

Load to install or dispatch a macOS GitHub Actions build, test or simulator screenshot run for an Xcode project.

skill
youtube-publish

Load to draft a YouTube upload, comment or reply, or send a draft after the user's recorded approval.

skill
block-argv-kill

Refuse a command that kills processes by name or by argv pattern (`pkill`, `killall`), whatever flags follow. On a machine where several agents share one checkout and its worktrees, a pattern that matches a tool's name matches every lane running that tool, the caller's own shell included when its command line holds the pattern. Names the accepted forms: `kill <pid>` on a PID the caller recorded wh

hook
block-bare-cd

Not run on codex: each shell call starts a fresh process. Not run on pi: each shell call starts a fresh process. Not run on antigravity: its toolCall.args payload is not read. Refuse a command with a line that is only a `cd`. Copilot's bash tool runs every call in a fresh process, so a bare cd passes there; its powershell tool keeps the check. A here-document body is not judged because it is input

hook
block-repo-copy

Block a copy (cp, rsync, tar, git clone) whose source names a `.git` or `target` path component and whose destination is under /tmp, /var/tmp or $TMPDIR. Suggests reading the source in place or building a minimal fixture.

hook
block-unsafe-rm

Block plain rm on shared directory roots, globs directly under them, child paths with . or .. segments, and paths that start with a variable that may expand empty. Keep a private mktemp directory and remove it in the same shell call. Not run on antigravity: its required command-safety companion does not name antigravity, whose payload is toolCall.args.

hook
claude-attention

Advisory reader for Claude permission prompts and the claude-question-attention companion. Requires Claude Code 2.1.246 or newer, Python 3.8+ and tmux. Uses native terminalSequence BEL only when no attached client views the source pane and a client attaches to its session, with a shared user-wide 10-second cooldown. Controls only its own emissions; consumers wanting one channel select preferredNot

hook
claude-question-attention

Advisory reader forwarding Claude question events to the installed claude-attention sibling, which owns admission, terminalSequence output and cooldown. Requires Claude Code 2.1.246 or newer, Python 3.8+ and tmux. Not run on codex: its question events do not provide Claude terminalSequence output. Not run on opencode: its question events do not provide Claude terminalSequence output. Not run on cu

hook
cloud-git-hooks

Runs the commit-guards installer and its read-only check only when CLAUDE_CODE_REMOTE=true and HEAD commits this hook's project SessionStart registration in .claude/settings.json. A global install reaches no repository without that committed registration. Reports an arming gap as session context without refusing startup. Not run on codex: its SessionStart environment supplies no CLAUDE_CODE_REMOTE

hook
command-safety

On harnesses that execute hooks, refuse supported relative recursive deletes, git init and .git output writes reachable after a failed cd in a flat literal list. Refuse cat reads of the known ~/.fleet/env spellings. Refuse a recognized whole-environment read where the shell runs it, unless a proven direct pipe filters named variables: at a command head, among the forwarded operands of a shell or s

hook
critical-path-deny

Answers Claude Code's critical-path check prompt for a Bash call with deny at once, inside a launched orch lane only, so an unattended lane is not held two minutes per prompt and does not trip the three-prompt circuit breaker. The deny message gives the rewrite the permission-modes page gives for each form the check flags: `${NAME:?}` or a literal path for a glob or trailing slash under a variable

hook
dev-stop-check

Blocks a subagent's stop while a validation run it started is still going, naming the `orch-runtime validate entry --wait --run-dir <run dir>` command to run next, because a subagent whose turn has ended is not woken when the run ends and its round is left with a verdict on disk and no commit. The worktrees are those the subagent's transcript names as `orch-runtime validate entry ... --worktree <a

hook
doc-drift-check

Retired: runs no check and exits 0. `kendex remove doc-drift-check --scope project --sweep` removes the declaration and its files in one change; a hand deletion is swept by the next rolling refresh. Not run on gemini: it has no Stop event. Not run on antigravity: its Stop payload carries no `stop_hook_active`.

hook
docs-writing-bash-note

After a successful Claude Code Bash call, runs the docs-writing-note advisory reader beside this hook to route the first changed tracked Markdown file once per session. Not run on codex: its apply_patch payload is read by docs-writing-note before the edit. Not run on pi: its edit payload is read by docs-writing-note before the edit. Not run on copilot: its postToolUse Bash payload is not read here

hook
docs-writing-note

On the first Markdown edit of a session, hands the session the edited file's docs-writing type and the paths of that type's reference and example, as the commit-guards md-type lane installed in this hook's own scope routes them, once per session. The first line is `docs-writing-note: path=<file> type=<type>`, or `type=unclassified` naming the route table in docs-writing's SKILL.md. Later Markdown

hook
jarvis-memory-note

Before an edit under a Jarvis home's memory directory, hands the session the edited path and the Notes reading route in the installed Jarvis memory guide. The home's installed jarvis/SKILL.md identifies the home. The guide alone owns the note rule. Not run on codex: apply_patch has no tool_input.file_path to identify the edited note. Not run on copilot: preToolUse has no measured additionalContext

hook
lane-mail-check

Holds a launched lane's turn end while its overseer's mail stands unread, at a context or account handoff mark, or when the lane reported nothing since its last judged turn end, and judges the fleet overseer's turn end at its own marks; the hooks that require it run it at their own events with an argument. The hook finds its own install, walking up from its directory to the home, the open reposito

hook
lane-mail-compact

Flags a Copilot lane's or overseer's automatic compaction, so its next turn end holds it until it writes its handoff record, the backstop for a turn that crossed into the compaction before its context reading reached a turn end. A Copilot session is handed off before its compaction by its context reading: the orch `copilot-lane-context` Copilot extension hands the lane-mail-check hook the tokens i

hook
lane-mail-deliver

Hands a lane the lines its overseer mailbox holds unread once a tool call finishes, so a directive reaches a working lane at its next tool call rather than at its turn end. The judgement is the lane-mail-check hook's, run from beside this one with the argument `deliver`: it exits 0 with `{"hookSpecificOutput":{"hookEventName":"PostToolUse","additionalContext":...}}` on stdout, the context opening

hook
lane-mail-halt

Refuses every tool call in a lane while its overseer mailbox holds an unread directive sent with `lane-mail send --halt`, so a halted lane stops at its next tool call. The judgement is the lane-mail-check hook's, run from beside this one with the argument `halt`: its refusal opens `lane-mail-check: halt=<id>` and carries the directive and the one `lane-mail inbox` command that reads it. That comma

hook
lane-mail-model-switch

Runs the installed lane-mail-check hook after Claude Code changes a launched lane or bound overseer session model. That hook owns model comparison, session records and recovery. Not run on codex: no PostModelSwitch event. Not run on pi: no PostModelSwitch event. Not run on copilot: no PostModelSwitch event. Not run on opencode: no PostModelSwitch event. Not run on cursor: no PostModelSwitch event.

hook
lane-mail-prompt

Hands a Claude Code, Codex or Copilot CLI lane the lines its overseer mailbox holds unread each time it is handed a prompt, so a lane idle at its prompt reads its mail with the next prompt rather than at its first tool call or turn end. The judgement is the lane-mail-check hook's, run from beside this one with the argument `prompt`: it exits 0 with context on stdout, opening `lane-mail-check: unre

hook
lane-mail-start

Hands a Copilot lane the lines its overseer mailbox holds unread when its session starts, so a lane launched or resumed onto waiting mail reads it before its first tool call rather than at its first tool call or turn end. The judgement is the lane-mail-check hook's, run from beside this one with the argument `start`: it exits 0 with `{"additionalContext":...}` on stdout, the context opening `lane-

hook
pre-commit-check

Defers commits to executable, marked pre-commit and commit-msg hooks in the working repository. Refuses a literal bypass option or core.hooksPath override on a direct git commit call. Also refuses the commit's own literal bypass flag after a core.hooksPath write or unset in the same command, and names the separate-command remedy. Reads quoted words, comments, command boundaries and option values w

hook
reviewer-read-only

For a subagent whose agent_type starts with `reviewer-`, refuses every Edit, MultiEdit and NotebookEdit call; a Write whose path lies inside a git work tree unless it is the review artifact, `<dir>/tmp/review-*.json`; and a Bash command that runs `git commit`, `push`, `checkout`, `restore`, `stash` (all subcommands), `clean`, `reset` or `switch` (options between `git` and the verb allowed). Its de

hook
reviewer-stop-check

Blocks a reviewer subagent's stop once when it leaves no readable review artifact or leaves files behind in the worktree it reviewed. The worktree is the one the artifact path in the subagent's transcript names (`<worktree>/tmp/review-<agent>-*.json`, the newest mention), the transcript being the payload's `agent_transcript_path` where the payload carries that key (Claude Code and Codex, whose `tr

hook
session-drift-check

On a fresh session start (not resume or compact), runs `kendex check --quiet --report-only` and surfaces kendex drift to the agent — outdated items (`kendex refresh`), items removed upstream (`kendex remove --kind <kind> <name>`, `--global` in a global section), unreachable sources, and packages not yet evaluated against their sources (a background refresh settles them). Outside a lane, prints not

hook
session-end-row

Records that a session ended, as one JSON row the fleet's overseer judgement reads instead of the session's pane: the runtime watch judges an overseer whose last row is a SessionEnd for any reason but `clear` or `resume` as exited only where its pane's process is a bare shell with nothing under it, and as live over a running process, whatever its screen shows. The row is written by the lane-mail-c

hook
session-start-row

Records that a session started, as one JSON row the fleet's overseer judgement reads instead of the session's pane. With TMUX set and TMUX_PANE absent, the orch runtime's turn end takes the pane from the live controller record only when its native session id matches. A session neither identity names receives one visible session-unjudged=missing-pane notice; row and Stop share that notice, and late

hook
stop-failure-row

Records that a turn ended on an API error, as one JSON row the fleet's overseer judgement reads instead of the session's pane: the runtime watch judges an overseer whose last row is a StopFailure with error `rate_limit` as walled, Claude Code's own word for a usage limit, or on Pi, which names no error kind, one whose `message` agent-runtime's limit-banner reader reads as the account's limit, what

hook
task-completed-check

Before a task is marked complete, runs `cargo clippy --all-targets` once for the workspace members that own the changed Rust files and every member depending on one of them, however indirectly, through a path dependency whose directory is that member's, `-p` per member as the one `cargo metadata --no-deps` read names them, against the repository's Cargo.toml, or the nearest one above a changed fil

hook
ui-copy-note

Gives the session the installed code-quality UI copy reference after the first matching edit attempt. Successful Claude Code Bash calls check tracked changes across the checkout, including changes already present. Matches each path relative to its project against QA_UI_PATHS. When the setting is unset or empty, the fallback is *.qml *.tsx *.jsx *.vue *.svelte. Codex apply_patch renames use the des

hook
worktree-prompt-claim

Uses the worktree-session-claim hook's decision before each Claude prompt. In required-claim mode its keyed claim failure on stderr and exit 2 block the prompt before the model processes it; an established session's `maintenance` notice does not block. Without that mode it passes silently. The SessionStart hook requires this companion on Claude. Not run on codex: SessionStart already supports the

hook
worktree-session-claim

Claims the linked git worktree a session starts in when its private git directory carries the worktree skill's `kendex-issue` record. The worktree skill's `references/session-guard.md` defines the lease and its limits. It runs `worktree-session-guard claim <worktree root>` with the guard's owner ladder: `KENDEX_SESSION_OWNER`, else `HT_SESSION_OWNER`, else `USER`. A workflow can adopt that initial

hook
code-scrub

Audit merged pull requests over a user-specified number of days.

command