Find your next AI tool.
Search tools, agents, MCP servers, and skills. Open a review to check features, prices, and limits.
More filters
4 entries
01Define done before the agent repeatsGoal and verifier loopA loop pattern where the agent starts from an explicit goal, runs a bounded step, then uses tests, source checks, or review criteria as the verifier before continuing.02Pause before consequence-bearing actionsHuman approval gateA loop pattern that lets agents prepare changes, evidence, and next actions, but requires human approval before deploys, permission changes, data writes, or vendor-facing messages.03Limit repeats before cost or drift compoundsBudget and stop-condition loopA loop pattern that defines maximum attempts, token or time budgets, escalation criteria, and fallback behavior before the agent starts a long-running task.04Source → draft → review → receipt → resultsReviewed social publication queueRepeat channel-specific content work with an authorized account, versioned drafts, deduplication and explicit recovery from uncertain publication.