{"repo":"NousResearch/hermes-agent","public_id":"28056","display_label":"Agent verification guards against unverified tool claims","summary":"These items address a pattern where the agent asserts success or factual conclusions after tool failures or without verifying external state. The work spans detection, prompting guidance, runtime guards, and quality gates to prevent unsupported completion claims.","subsystem":"agent","counts":{"members":10,"open_issues":6,"open_issues_distinct":6,"open_prs":2,"closed_issues":0,"merged_prs":1,"attempt_prs":3},"landed_fix":{"status":"merged_to_main","basis":"structural","pr":50009},"resolution":{"state":"unresolved","by":null},"freshness":{"state":"verified","last_verified_at":"2026-06-21T14:25:43+00:00","is_stale":false},"grouping":{"backbone":"STRUCTURAL_PARTIAL","link_density":0.12888888888888891},"fix_edges":[{"issue_number":8049,"pr_number":50009,"tier":"verified","basis":"closing_ref","confidence":0.95,"recoverable_from_timeline":true}],"members":[{"kind":"issue","number":28056,"title":"Add in-run quality gates with bounded retry for cron/agent runs","url":null,"state":"open"},{"kind":"issue","number":45577,"title":"CLI/runtime: block unsupported 'STATUS: DONE / verified / deployed' summaries after non-executing tool turns","url":null,"state":"open"},{"kind":"issue","number":48214,"title":"LLM overconfidence on factual assertions — agent answers from memory instead of verifying system state","url":null,"state":"open"},{"kind":"issue","number":53072,"title":"Agent may claim GitHub repo creation/push succeeded after tool failures without verification","url":null,"state":"open"},{"kind":"issue","number":54722,"title":"Agent can report verified success or factual conclusions after failed or unverified tool evidence","url":null,"state":"open"},{"kind":"issue","number":58196,"title":"[Feature]: Verification enforcement -- verify-before-claim guidance + unsupported-completion-claim detector","url":null,"state":"open"},{"kind":"pr","number":45629,"title":"fix(agent): guard unsupported completion claims","url":null,"state":"closed","close_reason":{"category":"volume_quota","implication":"active"},"merged":false},{"kind":"pr","number":50009,"title":"fix(agent): guard finalize_turn cleanup chain so it never drops the response","url":null,"state":"merged","close_reason":{"category":"merged","implication":"superseded"},"merged":true,"is_fix_pr":true,"fix_tier":"verified","files_changed":["agent/turn_finalizer.py","tests/agent/test_turn_finalizer_cleanup_guard.py"]},{"kind":"pr","number":53388,"title":"fix(agent): flag unverified external state after tool failures","url":null,"state":"open","close_reason":{"refs":[{"kind":"commit","value":"953ee63382a9ae3cede6f05104a864a604f61506","number":0}],"category":"implemented_elsewhere","implication":"superseded"},"merged":false},{"kind":"pr","number":54738,"title":"fix(agent): add generic tool-verifier footer for unverified tool failures","url":null,"state":"open","close_reason":{"category":"superseded_by_pr","implication":"superseded"},"merged":false}],"trust":{"default_tier":"verified","includes_suggested":false}}