The full skill instructions and embedded helper script are in skills/pdf2plot.md at the root of this repository. Read that file now and follow its instructions.
other
2026-10-11_00-33-53Z_EvolvAttack_BackdooringselfEvolvingVLMAgents
ANo problem foundflybfree/AI-Wiki
Self-evolving mechanisms let vision-language model (VLM) agents continually adapt through evolving memory, but they also allow malicious behavior to emerge only after deployment. We introduce a new attack surface, the self-evolving backdoor, whose payload stays dormant in the initial agent and is progressively activated as the agent evolves. Unlike conventional backdoors, which can be exposed by testing triggered inputs on a static model, our attack requires both a visual trigger and a sufficie…
2026-10-11_00-33-53Z_EvolvAttack_BackdooringselfEvolvingVLMAgents got grade A on our automated scan on 2026-10-23.
Before you install
- Read any scripts this skill ships.
- Note every URL it fetches.
- Check allowed-tools and hooks.
Embed this grade
Add this badge to your README. It shows the current grade from CheckSkills.
Report this skill if the grade looks wrong, or if something slipped through.
Related
Run all code generation (protobuf, sqlc, frontend build)
name: ship description: Release protocol: pre-flight checks, changelog, semantic version, tag, deploy. disable-model-invocation: true
name: init description: Bootstrap the .resonance/ project memory (soul, state, docs scaffold). Run once per new project. disable-model-invocation: true
name: goal description: The autonomous goal loop: frame, decompose, then build and verify each slice against real checks, bounded, never auto-ship. disable-model-invocation: true