improve agent CI matrix, token permissions, and waitlist follower backfill (#313)
* add workflows permission to git token and waitlist improvements - add `workflows` to `InstallationTokenPermissions` type in both action and API token routes - include `workflows: write` in the git token so agents can push workflow file changes - add `githubFollowers` field to WaitlistSignup schema with migration - add script to populate waitlist followers from GitHub API - add frog-green-square-border logo asset Co-authored-by: Cursor <cursoragent@cursor.com> * improve CI, agent logging, token permissions, and delegation guardrails - add format check and build step to root CI job - standardize agent model/effort log lines across all agents - fix GitHub App permissions types to match OpenAPI schema (workflows is write-only) - improve delegation error message to prevent subagent recursion - demote noisy OpenCode stderr to debug level - add subagent delegation rules to resolved instructions Co-authored-by: Cursor <cursoragent@cursor.com> * fix graphql partial error handling, update delegation message, add workflow_run fixtures Co-authored-by: Cursor <cursoragent@cursor.com> * remove module-level env var throws that break CI build Co-authored-by: Cursor <cursoragent@cursor.com> * fix logging bug and type hole from PR review - use batch-local notFound counter so per-batch log doesn't undercount - add workflows to WorkflowTokenPermissions so wire type matches what action sends Co-authored-by: Cursor <cursoragent@cursor.com> * lazy-init appOctokit to fix next build without env vars Co-authored-by: Cursor <cursoragent@cursor.com> * drop pnpm build from CI test workflow Co-authored-by: Cursor <cursoragent@cursor.com> * fix delegate-effort test regex to match actual log format, disable fail-fast for agnostic tests the test was matching `running \w+ with effort=auto` but the actual log line from shared.ts is `» effort: auto`. also temporarily set fail-fast: false on action-agnostic so all failures surface at once. Co-authored-by: Cursor <cursoragent@cursor.com> * disable fail-fast in action workflow too, relax ci.test.ts to match both workflow files now use fail-fast: false for agnostic tests so all matrix jobs run to completion. the ci consistency test now checks that the two workflows agree rather than requiring true. Co-authored-by: Cursor <cursoragent@cursor.com> * restore fail-fast: true now that all agnostic tests pass Co-authored-by: Cursor <cursoragent@cursor.com> * skip agent tests in CI when agent harness file didn't change adds action/test/changed-agents.sh which reads the PR diff (via dorny/paths-filter) and outputs only agents whose harness file was modified. the action-agents matrix now uses this dynamic list instead of a hardcoded array, so e.g. a PR touching only cursor.ts runs 6 jobs instead of 30. Co-authored-by: Cursor <cursoragent@cursor.com> * update ci.test.ts to validate dynamic agent matrix the test now checks that the matrix references the changes job output and that changed-agents.sh correctly discovers all agents. Co-authored-by: Cursor <cursoragent@cursor.com> * parallelize action jobs and use claude canary fallback for shared changes runs action-agents in parallel with action-agnostic after root/changes, and updates changed-agents logic so shared or non-harness action runtime changes run only claude while harness-specific edits run only those changed agents. Co-authored-by: Cursor <cursoragent@cursor.com> --------- Co-authored-by: Cursor <cursoragent@cursor.com>
This commit is contained in:
Colin McDonnell
committed by
pullfrog[bot]
pullfrog[bot]
parent
37dcea86b9
commit
b6658ddbc1
+1
-1
@@ -86,7 +86,7 @@ export const claude = agent({
|
||||
// select model and effort level
|
||||
const model = claudeEffortModels[ctx.payload.effort];
|
||||
const effortLevel = claudeEffortLevels[ctx.payload.effort];
|
||||
log.info(`» using model: ${model}${effortLevel ? ` (effort: ${effortLevel})` : ""}`);
|
||||
log.info(`» model: ${model}${effortLevel ? ` (effort: ${effortLevel})` : ""}`);
|
||||
|
||||
// build disallowedTools based on tool permissions
|
||||
const disallowedTools = buildDisallowedTools(ctx);
|
||||
|
||||
+3
-4
@@ -155,10 +155,9 @@ export const codex = agent({
|
||||
|
||||
// get model and reasoning effort based on effort level
|
||||
const effortConfig = getCodexEffortConfig(model)[ctx.payload.effort];
|
||||
log.info(`» using model: ${effortConfig.model} (effort: ${ctx.payload.effort})`);
|
||||
if (effortConfig.reasoningEffort) {
|
||||
log.info(`» using modelReasoningEffort: ${effortConfig.reasoningEffort}`);
|
||||
}
|
||||
log.info(
|
||||
`» model: ${effortConfig.model}${effortConfig.reasoningEffort ? ` (reasoningEffort: ${effortConfig.reasoningEffort})` : ""}`
|
||||
);
|
||||
|
||||
// determine sandbox mode based on push permission
|
||||
// push: "disabled" → read-only sandbox, otherwise workspace-write.
|
||||
|
||||
+5
-4
@@ -134,7 +134,7 @@ export const cursor = agent({
|
||||
try {
|
||||
const projectConfig = JSON.parse(readFileSync(projectCliConfigPath, "utf-8"));
|
||||
if (projectConfig.model) {
|
||||
log.info(`» using model from project .cursor/cli.json: ${projectConfig.model}`);
|
||||
log.info(`» model: ${projectConfig.model} (from .cursor/cli.json)`);
|
||||
} else {
|
||||
modelOverride = cursorEffortModels[ctx.payload.effort];
|
||||
}
|
||||
@@ -146,9 +146,9 @@ export const cursor = agent({
|
||||
}
|
||||
|
||||
if (modelOverride) {
|
||||
log.info(`» using model: ${modelOverride}, effort=${ctx.payload.effort}`);
|
||||
log.info(`» model: ${modelOverride}`);
|
||||
} else if (!existsSync(projectCliConfigPath)) {
|
||||
log.info(`» using default model, effort=${ctx.payload.effort}`);
|
||||
log.info(`» model: default`);
|
||||
}
|
||||
|
||||
// track logged model_call_ids to avoid duplicates
|
||||
@@ -439,5 +439,6 @@ function configureCursorTools(ctx: AgentRunContext): void {
|
||||
}
|
||||
|
||||
writeFileSync(cliConfigPath, JSON.stringify(config, null, 2), "utf-8");
|
||||
log.info(`» CLI config written to ${cliConfigPath}`, JSON.stringify(config, null, 2));
|
||||
log.info(`» CLI config written to ${cliConfigPath}`);
|
||||
log.debug(`» CLI config contents: ${JSON.stringify(config, null, 2)}`);
|
||||
}
|
||||
|
||||
+1
-1
@@ -345,7 +345,7 @@ function configureGeminiSettings(ctx: AgentRunContext): string {
|
||||
// allow env var override for tests (e.g., to avoid flash RPD quota limits)
|
||||
const model = process.env.GEMINI_MODEL ?? effortConfig.model;
|
||||
const thinkingLevel = effortConfig.thinkingLevel;
|
||||
log.info(`» using model: ${model}, thinkingLevel: ${thinkingLevel}`);
|
||||
log.info(`» model: ${model} (thinkingLevel: ${thinkingLevel})`);
|
||||
|
||||
const realHome = homedir();
|
||||
const geminiConfigDir = join(realHome, ".gemini");
|
||||
|
||||
+25
-25
@@ -72,7 +72,9 @@ export const opencode = agent({
|
||||
const modelOverride = process.env.OPENCODE_MODEL;
|
||||
if (modelOverride) {
|
||||
args.push("--model", modelOverride);
|
||||
log.info(`» using model override: ${modelOverride}`);
|
||||
log.info(`» model: ${modelOverride} (override)`);
|
||||
} else {
|
||||
log.info(`» model: auto-selected by OpenCode`);
|
||||
}
|
||||
|
||||
process.env.HOME = tempHome;
|
||||
@@ -186,13 +188,10 @@ export const opencode = agent({
|
||||
lastProviderError = providerError;
|
||||
log.error(`» provider error detected (${providerError}): ${trimmed.substring(0, 500)}`);
|
||||
} else {
|
||||
// try to parse as JSON for structured logging, fall back to warning
|
||||
try {
|
||||
const parsed = JSON.parse(trimmed);
|
||||
log.debug(JSON.stringify(parsed, null, 2));
|
||||
} catch {
|
||||
log.warning(trimmed);
|
||||
}
|
||||
// OpenCode's --print-logs output goes to stderr. demote internal
|
||||
// INFO/DEBUG bus traffic to debug so it doesn't drown out tool
|
||||
// call logs in the GitHub Actions step output.
|
||||
log.debug(trimmed);
|
||||
}
|
||||
},
|
||||
});
|
||||
@@ -325,7 +324,7 @@ function configureOpenCode(ctx: AgentRunContext): void {
|
||||
}
|
||||
|
||||
log.info(`» OpenCode config written to ${configPath}`);
|
||||
log.info(
|
||||
log.debug(
|
||||
`» OpenCode permissions: edit=${permission.edit}, bash=${permission.bash}, webfetch=${permission.webfetch}`
|
||||
);
|
||||
log.debug(`OpenCode config contents:\n${configJson}`);
|
||||
@@ -558,27 +557,28 @@ const messageHandlers = {
|
||||
const status = event.part?.state?.status;
|
||||
const output = event.part?.state?.output;
|
||||
|
||||
// debug log all tool_use events to diagnose missing bash/MCP tool calls
|
||||
if (!toolName || !toolId) {
|
||||
log.debug(`» tool_use event missing toolName or toolId: ${JSON.stringify(event)}`);
|
||||
// surface dropped tool_use events visibly so missing tool calls are diagnosable
|
||||
log.info(
|
||||
`» tool_use event missing toolName or toolId: ${JSON.stringify(event).substring(0, 500)}`
|
||||
);
|
||||
return;
|
||||
}
|
||||
|
||||
if (toolName && toolId) {
|
||||
// track tool call in current step
|
||||
if (stepHistory.length > 0) {
|
||||
stepHistory[stepHistory.length - 1].toolCalls.push(toolName);
|
||||
}
|
||||
// track tool call in current step
|
||||
if (stepHistory.length > 0) {
|
||||
stepHistory[stepHistory.length - 1].toolCalls.push(toolName);
|
||||
}
|
||||
|
||||
thinkingTimer.markToolCall();
|
||||
log.toolCall({
|
||||
toolName,
|
||||
input: parameters || {},
|
||||
});
|
||||
thinkingTimer.markToolCall();
|
||||
log.toolCall({
|
||||
toolName,
|
||||
input: parameters || {},
|
||||
});
|
||||
|
||||
// if tool already completed (status in same event), log output
|
||||
if (status === "completed" && output) {
|
||||
log.debug(` output: ${output}`);
|
||||
}
|
||||
// if tool already completed (status in same event), log output
|
||||
if (status === "completed" && output) {
|
||||
log.debug(` output: ${output}`);
|
||||
}
|
||||
},
|
||||
tool_result: (event: OpenCodeToolResultEvent, thinkingTimer: ThinkingTimer) => {
|
||||
|
||||
+10
-8
@@ -28,13 +28,15 @@ export const agent = <const input extends AgentInput>(input: input): defineAgent
|
||||
return {
|
||||
...input,
|
||||
run: async (ctx: AgentRunContext): Promise<AgentResult> => {
|
||||
const bash = ctx.payload.bash;
|
||||
const web = ctx.payload.web;
|
||||
const search = ctx.payload.search;
|
||||
const push = ctx.payload.push;
|
||||
log.info(
|
||||
`» running ${input.name} with effort=${ctx.payload.effort}, timeout=${ctx.payload.timeout}...`
|
||||
);
|
||||
log.info(`» agent: ${input.name}`);
|
||||
log.info(`» effort: ${ctx.payload.effort}`);
|
||||
if (ctx.payload.timeout) log.info(`» timeout: ${ctx.payload.timeout}`);
|
||||
log.info(`» web: ${ctx.payload.web}`);
|
||||
log.info(`» search: ${ctx.payload.search}`);
|
||||
log.info(`» push: ${ctx.payload.push}`);
|
||||
log.info(`» bash: ${ctx.payload.bash}`);
|
||||
log.debug(`» payload: ${JSON.stringify(ctx.payload, null, 2)}`);
|
||||
|
||||
// build log box content: eventInstructions (if any) + user request (if any) + event data
|
||||
const logParts = [
|
||||
ctx.instructions.eventInstructions
|
||||
@@ -46,7 +48,7 @@ export const agent = <const input extends AgentInput>(input: input): defineAgent
|
||||
log.box(logParts.join("\n\n---\n\n"), {
|
||||
title: "Instructions",
|
||||
});
|
||||
log.info(`» tool permissions: web=${web}, search=${search}, push=${push}, bash=${bash}`);
|
||||
|
||||
return input.run(ctx);
|
||||
},
|
||||
...agentsManifest[input.name],
|
||||
|
||||
Reference in New Issue
Block a user