Open-source agentic AI Search Execution Platform. Track AI visibility, find opportunities, and let agents turn signals into prioritized actions and drafted fixes across ChatGPT, Claude, Gemini, Google AI Overviews, Google AI Mode, Perplexity, Grok, and Microsoft Copilot.
TypeScript aeoai-agentai-searchanswer-engine-optimizationbrand-monitoringchatgptclaudecopilotdockergeminigeogoogle-ai-overviewsgrokmcpnextjsopen-sourceperplexityseosupabasetypescript
-- Task execution modes and permission levels, in the specification's terms.
--
-- Phase 5 (00088) stored a simplification: `mode` as manual/agent and
-- `permission` as none/approval. The V1 Task Library specification defines
-- four execution modes and three permission levels, and the difference is not
-- cosmetic — SYSTEM work (a deterministic comparison) and HUMAN_OR_AGENT work
-- (either may do it, depending on the tools connected) are both things the
-- planner and the runner have to be able to say, and manual/agent cannot.
--
-- mode system | agent | human | human_or_agent
-- permission read | write | execute
--
-- WRITE and EXECUTE require approval unless a workspace explicitly authorises
-- autonomy; READ does not. That is the specification's default and what the
-- runner enforces.
--
-- The constraints are dropped before the rows are rewritten: 00083 rewrote
-- rows first and the old constraint rejected them, rolling the migration back.
alter table public.action_tasks drop constraint if exists action_tasks_mode_check;
alter table public.action_tasks drop constraint if exists action_tasks_permission_check;
update public.action_tasks
set mode = case mode when 'manual' then 'human' else mode end,
permission = case permission when 'none' then 'read' when 'approval' then 'execute' else permission end;
alter table public.action_tasks alter column mode set default 'human';
alter table public.action_tasks alter column permission set default 'read';
alter table public.action_tasks
add constraint action_tasks_mode_check
check (mode in ('system', 'agent', 'human', 'human_or_agent'));
alter table public.action_tasks
add constraint action_tasks_permission_check
check (permission in ('read', 'write', 'execute'));
comment on column public.action_tasks.mode is
'Who carries the task out: system (deterministic, by Ansvisor), agent, human, or human_or_agent.';
comment on column public.action_tasks.permission is
'read, write (modifies owned content) or execute (acts externally). write and execute require approval unless autonomy is authorised.';