Sales Oratory Master

Handles B2B presales and renewal negotiations with value-shift tactics and compliance checks. Use when the user mentions customer pushback on price, budget c...

This skill appears coherent and limited to producing compliant sales talktracks via a supplied LLM client. Before installing: 1) Review where your agent’s llm_client sends data — customer quotes and prompts will be transmitted to that model endpoint, so confirm data-handling and privacy with your provider. 2) Because the package owner and homepage are unknown, consider running the included tests in an isolated environment to verify behavior. 3) The PROMISE_GUARD enforces helpful redlines, but also manually review outputs for legal/industry compliance in your jurisdiction. 4) Avoid sending sensitive PII or contract terms to third-party LLM endpoints unless you’ve approved that data sharing. If those checks are acceptable, the skill is proportionate to its stated purpose.

Type: OpenClaw Skill Name: sales-oratory-master Version: 1.0.0 The skill is vulnerable to prompt injection. In `main.py`, user-provided inputs (`customer_quote`, `deal_stage`, `client_persona`) are directly embedded into the `prompt_template.md` using string formatting, which then forms part of the `system_message` sent to the LLM. A malicious user could craft these inputs to manipulate the LLM's behavior, potentially overriding the `PROMISE_GUARD.md` instructions or causing the LLM to generate unintended or harmful content. This is a significant vulnerability, but there is no evidence of intentional malicious code (e.g., data exfiltration, backdoors) within the skill itself.