tick-md
PassAudited by ClawScan on May 1, 2026.
Overview
The skill is coherent and shows no malicious behavior, but it relies on external Tick CLI/MCP tooling and lets agents persistently modify project task records.
Before installing, make sure you trust the Tick npm packages and want an MCP server available in your editor. Approve config edits and git pushes manually, keep TICK.md free of secrets, and use the autonomous worker role only when you are comfortable with agents claiming and updating project tasks.
Findings (5)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Installing the skill may lead users to install command-line and MCP tooling that can act in their project workspace.
The skill relies on external globally installed npm packages and an MCP server rather than bundled reviewed code; this is expected for the integration but users should trust the package source.
npm install -g tick-md ... npm install -g tick-mcp-server
Install only from trusted package sources, review package names carefully, and keep the tool updated through normal package-management practices.
If approved, the assistant can use Tick MCP tools to read and modify the project's task coordination file.
Adding the Tick MCP server changes the assistant's available tool surface in the editor; the same guide also instructs that config changes require explicit user approval.
Add this to your editor's MCP config file
Manually review and approve MCP configuration changes, back up config files first, and remove the MCP entry if you no longer want the tools available.
An agent could change, delete, or reorder task status in TICK.md if asked or if using the coordination workflow.
The documented tools can directly modify task state, including operations that bypass normal workflow validation; this is disclosed as a correction feature but can affect shared project coordination.
Directly edit task fields, bypassing state machine validation.
Use dry-runs and undo where available, review destructive commands such as delete/force, and keep remote pushes approval-gated as the docs recommend.
Task comments or stale entries in TICK.md may influence later agent behavior during coordination.
The skill intentionally stores task state, comments, and history in a persistent Markdown file that future agents may read and act on.
maintain tasks transparently in TICK.md
Treat TICK.md as shared project state, avoid placing secrets in task comments, and have agents validate task instructions against current user intent.
A worker-style agent may claim available tasks and continue coordination work across a session without asking about every local task-state change.
The worker role documents an autonomous check-in loop for claiming and continuing tasks; this is disclosed and purpose-aligned, but users should bound when such behavior is allowed.
worker agent that polls for available tasks and works autonomously
Use worker mode only in repositories where autonomous task claiming is desired, and set clear session boundaries for what the agent may work on.