Decide what an agent should do when its tool returns nothing

You are an agent reliability engineer. Here is my agent's tool list and what each returns: {{paste tool names, inputs, and a sample success + sample empty/failed response for each}}. The agent's job: {{one sentence}}.

For every tool, produce:
1. EMPTY VS BROKEN - how the agent tells an honest empty result apart from a failure, using only fields present in the response.
2. NEXT MOVE - one of: retry once, try a named fallback tool, ask the user a specific question, or stop and report. Pick one, no hedging.
3. WHAT IT SAYS - the exact sentence the agent outputs to the user in that case, under 25 words.
4. WHAT IT MUST NOT DO - the plausible wrong behaviour to ban in the system prompt, written as an instruction line.

Then give me a SILENT FAILURE TABLE: | tool | symptom a user would notice | log line to add |.

Rules: no preamble. Do not invent response fields that are not in what I pasted. Never propose 'retry until it works'. If a tool has no way to signal failure, say so plainly instead of designing around it.