{"repo":"NousResearch/hermes-agent","public_id":"54256","display_label":"Skill instructions not enforced as mandatory by agent","summary":"A reported bug and related fix attempts address agents treating injected skill instructions as optional or advisory rather than mandatory. Multiple fix iterations suggest ongoing work to make skill-loading prompts more directive.","subsystem":"skills","counts":{"members":2,"open_issues":1,"open_issues_distinct":1,"open_prs":0,"closed_issues":1,"merged_prs":0,"attempt_prs":0},"landed_fix":null,"resolution":{"state":"unresolved","by":null},"freshness":{"state":"never_verified","last_verified_at":null,"is_stale":false},"grouping":{"backbone":"LLM_ONLY","link_density":0.36},"fix_edges":[],"members":[{"kind":"issue","number":54256,"title":"[Bug]: Skill System Degradation — Injected Skills Treated as \"Advisory Context\" Instead of Mandatory Instructions Under Prompt Pollution","url":null,"state":"closed"},{"kind":"issue","number":58569,"title":"[Bug]: Agent loads skills but doesn't follow them — skill instructions treated as advisory","url":null,"state":"open"}],"trust":{"default_tier":"verified","includes_suggested":false}}