MCP server for Mock Screenshots — generate deep links to pixel-perfect fake chat & screenshot generators (iMessage, WhatsApp, Instagram, Telegram, Messenger, Snapchat).
Three tools with complete, well-structured schemas and descriptions. Tool naming follows verb_noun convention (generate_fake_chat, list_platforms, list_devices). Descriptions are clear and actionable (194 - 256 chars, within baseline 34 - 392 range). Input schemas are comprehensive with proper enums, constraints, and field descriptions. generate_fake_chat includes detailed parameter documentation with privacy warnings and validation rules. Error handling is present but generic. No tool annotations (readOnlyHint/destructiveHint) despite all three being read-only operations. Output schemas are not explicitly documented in the source.
Compose a fake chat screenshot (iMessage, WhatsApp, Instagram DM, Telegram, Messenger, Snapchat) and get a deep link to the Mock Screenshots generator, pre-filled and ready to preview and download. Output is watermarked and clearly fictional; intended for parody, education, design mockups and fiction — not deception.
List the device frames supported by the generator.
List the chat platforms supported by the generator.
Tool annotations missing: all three tools are read-only but lack readOnlyHint in schema. Agents cannot infer safety from tool definitions alone.
Output schemas not documented in source code. LLMs cannot plan downstream calls or extract fields without knowing response structure.
list_platforms and list_devices have minimal descriptions (under 20 chars would trigger hard cap, but these are ~70 chars). Descriptions lack context on when to call these discovery tools or what structure they reveal.
Error handling returns generic isError flag with text content. No error classification (retryable vs user-fixable vs fatal) or recovery guidance beyond ethics warning.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 62 | 2026-07-28+ | v2 |
generate_fake_chat accepts base64-encoded image data as parameter. Privacy warning is present but embedded in code; agents may not see it. No guidance on treating attachment URLs as sensitive in transcripts.