pulls/pr_01m47d24b0e6n91zwymwxg0vpx

public

32 lines1,157 bytesCode Blame

Pick any line to see why it is the way it is: the commit, the pull request and issue it came from, and what the agent was thinking.

Agents get guardrails, run credentials, an audit log, a context hub, repository instructions and mentions; security upkeep; snake_case API2d ago	1	-- Comments that mention @g1t-agent, each handled once: the runner takes a
	2	-- pending one when it hears the comment was made, and says what it did.
	3	CREATE TABLE agent_mentions (
	4	comment_id TEXT PRIMARY KEY,
	5	repo_id TEXT NOT NULL,
	6	number INTEGER NOT NULL,
	7	-- Set when the comment is on a pull request.
	8	pull_id TEXT,
	9	-- Who wrote it, as JSON, with the memberships they had when they did.
	10	actor TEXT NOT NULL,
	11	body TEXT NOT NULL,
	12	-- What they asked for: work, question or review.
	13	intent TEXT NOT NULL,
	14	-- Whether they belong to the repository's workspace.
	15	member INTEGER NOT NULL,
	16	-- pending, taken, then replied once g1t-agent has answered in the thread.
	17	status TEXT NOT NULL DEFAULT 'pending',
	18	outcome TEXT,
	19	created_at TEXT NOT NULL,
	20	taken_at TEXT
	21	);
	22	CREATE INDEX agent_mentions_by_subject ON agent_mentions (repo_id, number);
	23
	24	-- Rules that put g1t-agent to work without anyone assigning it. One row per
	25	-- repository that has set one.
	26	CREATE TABLE agent_rules (
	27	repo_id TEXT PRIMARY KEY,
	28	-- When an issue is given this label, g1t-agent takes it.
	29	label TEXT,
	30	updated_by TEXT NOT NULL,
	31	updated_at TEXT NOT NULL
	32	);