Job application tracker and AI-powered job search assistant. Helps job seekers manage their search journey with AI resume review, job matching, task logging, and application analytics, all while keeping your data private.
TypeScript ai-agentai-resume-analyzerapplication-trackerapplication-tracking-systematscareerjob-applicationjob-automationjob-huntingjob-searchjob-seeker-toolsjob-trackerjob-trackingjobsearchjobseekersmcpollamaresumeresume-reviewself-hosted
import "server-only";
import { tool, type LanguageModel, type UIMessageStreamWriter } from "ai";
import { TEMPERATURES } from "@/lib/ai/config";
import { APP_CONSTANTS } from "@/lib/constants";
import { AGENT_TOOL_DESCRIPTIONS } from "@/lib/agent/prompt";
import { AgentReviewResumeSchema } from "@/models/agent.schema";
import { resolveResumeForAgent } from "@/lib/agent/resumeLookup";
import { preprocessResume } from "@/lib/ai/tools/preprocessing";
import {
RESUME_REVIEW_SYSTEM_PROMPT,
buildResumeReviewPrompt,
} from "@/lib/ai/prompts/resume-review";
import { parseResumeReview } from "@/lib/ai/resumeReview/parse";
import { saveResumeReviewResult } from "@/actions/profile.actions";
import {
runNestedGeneration,
type NestedGenerationGuard,
} from "@/lib/agent/nestedGeneration";
import type { AgentReviewResumeResult } from "@/models/agent.model";
import type { ResumeReviewData } from "@/models/ai.schemas";
type ReviewResumeContext = {
userId: string;
pageResumeId?: string;
model: LanguageModel;
provider: string;
modelName: string;
writer: UIMessageStreamWriter;
guard: NestedGenerationGuard;
};
/**
* A second generation surface wearing a tool's clothes. execute makes its own
* streamText call with the dedicated review route's exact inputs, so parity is
* true by construction rather than by two prompts agreeing. userId and
* pageResumeId are closure parameters — the same IDOR boundary as the other
* tools, and it has a dedicated test.
*/
export function buildReviewResumeTool(ctx: ReviewResumeContext) {
return tool({
description: AGENT_TOOL_DESCRIPTIONS.review_resume,
inputSchema: AgentReviewResumeSchema,
// No needsApproval: this writes only the caller's own reviewData, and
// they watch the whole thing stream before it lands.
execute: async (raw, { toolCallId, abortSignal }): Promise<AgentReviewResumeResult> => {
const lookup = await resolveResumeForAgent(ctx.userId, {
title: raw?.resumeTitle,
pageResumeId: ctx.pageResumeId,
});
if (lookup.status === "no_resumes") return { status: "no_resumes" };
if (lookup.status === "needs_selection") {
return { status: "needs_selection", resumes: lookup.resumes };
}
const title = lookup.resume.title;
const pre = await preprocessResume(lookup.resume);
if (!pre.success) {
return {
status: "unreadable",
title,
reason: "It may be too short or missing content. Check it in Profile → Resumes.",
};
}
const generation = await runNestedGeneration({
model: ctx.model,
userId: ctx.userId,
system: RESUME_REVIEW_SYSTEM_PROMPT,
prompt: buildResumeReviewPrompt(pre.data.normalizedText),
temperature: TEMPERATURES.FEEDBACK,
numCtx: APP_CONSTANTS.AI_OLLAMA_NUM_CTX,
timeoutMs: APP_CONSTANTS.AI_RESUME_REVIEW_TIMEOUT_MS,
writer: ctx.writer,
toolCallId,
abortSignal,
guard: ctx.guard,
label: "review_resume",
provider: ctx.provider,
modelName: ctx.modelName,
attrs: {
"jobsync.input.resume_chars": pre.data.normalizedText.length,
},
});
if (generation.status === "busy") {
return {
status: "generation_failed",
title,
reason: "Another analysis is already running — ask for this one once it finishes.",
};
}
if (generation.status === "incomplete") {
return {
status: "generation_failed",
title,
reason: "The review stopped before it finished, so nothing was saved.",
};
}
if (generation.status === "failed") {
return {
status: "generation_failed",
title,
reason: "The review could not be generated. Try again in a moment.",
};
}
const { scores, body } = parseResumeReview(generation.text);
// reviewData is overwrite-only with no history, so an unscored or
// half-finished generation must not be written.
if (!scores) {
return {
status: "generation_failed",
title,
reason: "The review came back without a scores line, so nothing was saved.",
};
}
const reviewData: ResumeReviewData = {
...scores,
body,
reviewedAt: new Date().toISOString(),
provider: ctx.provider,
model: ctx.modelName,
surface: APP_CONSTANTS.AGENT_CHAT_REVIEW_SURFACE,
};
const saveResult = await saveResumeReviewResult(
lookup.resume.id!,
JSON.stringify(reviewData),
);
const saved = saveResult?.success === true;
return {
status: "ok",
resumeId: lookup.resume.id!,
title,
scores,
body,
saved,
saveError: saved ? undefined : (saveResult?.message ?? "The review could not be saved."),
};
},
});
}