{"found":true,"collection":"failure_modes","library_url":"https://www.grokbotdb.com/library/fm.test-run-does-real-work","generated":"2026-09-06T05:32:11Z","citation":["Quote the record id and its source_url (or provenance if there is no URL).","Include last_verified or observed_at when present. Do not present an old date as current.","If tier is community and confidence is low, say so. Do not launder it into a fact.","If the database does not hold the answer, say that. Do not invent Grok Bot product claims.","Contradictions are data: when conflicts_with is present, report both records rather than picking a favourite.","Deprecated and removed records stay visible. They are history, not current guidance."],"record":{"id":"fm.test-run-does-real-work","title":"There is no dry run, so a test destroys or publishes real things","symptom":"A run labelled as a test navigates real websites, changes real files and calls real connected tools.","cause":"The product has no rehearsal or sandbox mode. Approvals gate only the proposed action and do not reverse work already completed, so safety is entirely preventive.","detection":"Stated plainly in the official documentation and repeated by five independent third-party sources; not something a user discovers, something they must be told.","fix":"Test only on safe inputs. Gate every external write behind approval from the first run, using Allow once during testing and Always Allow only after a workflow is stable. Require approval for sending, purchasing, deleting, publishing and changing production systems. One source proposes building an application-level dry-run output contract into the skill itself, since the product provides none.","severity":"high","source_id":"src.tp.developersdigest.routines","source_url":"https://www.developersdigest.tech/blog/grok-bot-routines-automations-guide","confidence":"high","first_seen":"2026-09-03","last_verified":"2026-09-03"}}