AI agents use completeTask to create or update resources in Ava — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Ava environment.
The tool performs a write operation—it updates task status to 'complete' and triggers a Slack notification. While it modifies data, the action is reversible (a completed task can be reopened or corrected via updateTask or resumeTask). The severity is medium because misuse could cause workflow confusion or false completion claims, but the impact is limited in scope and remediable.
From the tool's definition The tool name 'completeTask' and description referencing '完了報告を共有するための入力仕様' (input specification for sharing completion reports) indicate this modifies task state by marking tasks as complete and shares information to Slack.
Documented attack patterns abuse exactly the kind of access completeTask gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Ava, and nothing reaches the server without passing your rules. This is the rule we recommend for completeTask:
{
"version": "1",
"default": "deny",
"tools": {
"completeTask": {
"limits": [
{
"counter": "completetask_rate",
"window": "minute",
"max": 30,
"scope": "grant"
}
]
}
}
}completeTask stays usable, but capped — an agent stuck in a loop can't make hundreds of changes a minute. Everything else on the server is denied unless you say otherwise.
Free to start. No card required.
完了報告を共有するための入力仕様。. It is categorised as a Write tool in the Ava MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Ava MCP server in PolicyLayer and add a rule for completeTask: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Ava. Nothing to install.
completeTask is a Write tool with medium risk. Write tools should be rate-limited to prevent accidental bulk modifications.
Yes. Add a rate_limit block to the completeTask rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for completeTask. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
completeTask is provided by the Ava MCP server (yutakobayashidev/ava). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Ava, add the rest of your stack, and see everything your agents can call. Then put policy on all of it.
Free to start. No card required.
10 Ava tools catalogued and risk-classified — across an index of 43,000+ MCP servers.