autoapply-mcp
Installed, started with no arguments, no environment and no input, and listed its tools. Run 2026-11-16.
Needs to start: nothing
Returns a result when called with no setup: 1 of 5 tools tried
Contacts while installing and starting: nothing
What it contacted
Nothing besides the npm registry it was installed from. Watched from outside its sandbox while it was installed and started with no input. Its tools may still contact things when called.
Starting it
npx -y -p autoapply-mcp@0.3.4 autoapply-mcp
Installing a package runs its code. Use a sandbox for anything you have not read.
Calling its tools
5 tools that take no input and look read-only were called once with nothing. 1 returned a result.
list_queue | No usable answer tool reported an error: [config_missing] profile not found at /tmp/.autoapply/profile.json. See docs/CONFIGURATION.md or run the "doctor" script. |
campaign_status | No usable answer tool reported an error: [config_missing] profile not found at /tmp/.autoapply/profile.json. See docs/CONFIGURATION.md or run the "doctor" script. |
list_sources | Returned a result under 0.1 s · 8,774 characters |
list_batches | No usable answer tool reported an error: [config_missing] profile not found at /tmp/.autoapply/profile.json. See docs/CONFIGURATION.md or run the "doctor" script. |
list_rules | No usable answer tool reported an error: [config_missing] profile not found at /tmp/.autoapply/profile.json. See docs/CONFIGURATION.md or run the "doctor" script. |
During the calls (not necessarily because of them) it first contacted: spanishtoptechcompanies.com.
Only tools whose names read like looking something up, with no required input, are tried, five at most. Tools that need input, and anything that sounds like it changes or sends something, are not called.
What its tools tell an agent
In the tool descriptions, nothing that talks about the user, other tools or private files, or uses hidden characters.
Found by simple pattern rules in each tool's description (up to 3,000 characters), quoted as written. Parameter descriptions are not scanned. A clean result is not a safety review.
Tools
discover_jobs companies, includeIssues
Fetches every configured ATS board, normalizes the postings, applies the campaign's hard gates, scores what survives, and stores the results. Read-only against employers: no application is created or sent.
resolve_company_board companyName, extraSlugs
Probes Greenhouse, Lever and Ashby for a company's public board slug. Board tokens are not published centrally, so guesses must be verified before being added to companies.json. Workday is not probed: its tenant, datacenter and site slug cannot be guessed, so find that URL on the employer's careers page and record it with add_company_board.
add_company_board name, ats, board, tier, tags, query, region, save
Checks that a candidate board actually serves postings, then saves it to companies.json. Nothing is saved unless postings were seen, because a wrong slug often answers 200 with a generic page rather than an error. Use this to record boards found elsewhere — a careers-page URL or a web search — including Workday, which cannot be probed by guessing.
scan_hiring_thread threadId, save, tier, limit
Scans a monthly Hacker News 'Ask HN: Who is hiring?' thread for links to employers' own ATS boards, verifies each one actually serves postings, and optionally saves the new ones to companies.json. This is a lead source, not a job source: the thread's prose cannot be gated on reliably, so only the board slug is taken and discover_jobs then ingests every role at that company with full structured pay and location. Free to call — it uses the public Hacker News API and no third-party scraping credits.
list_queue minScore, tiers, trackId, trackIds, locationClasses, companies, minCompensation, allowUnknownCompensation, limit
Returns gated-in jobs ordered by score. Roles already applied to are excluded by fingerprint, so the same posting on a second board is not surfaced twice.
29 more tools
explain_job jobId, includeDescription
Returns the full gate result, score breakdown with quoted evidence, and the posting text wrapped as untrusted data.
campaign_status
Summarizes progress toward the application target, pipeline health, tier distribution and the most common rejection reasons.
list_sources
Lists built-in ATS adapters, public lead APIs and portfolio pages, with access requirements and submission limitations. Never reveals API keys or grants company permission.
search_job_sources query, country, location, company, limit, page, source
Searches one bounded Himalayas or Foorilla page. Returns unverified leads with provenance, salary units and attribution, not stored employer jobs or application permission. Himalayas uses country; Foorilla requires AUTOAPPLY_FOORILLA_API_KEY and uses location.
scan_source_page source, url, limit
Reads one a16z, Sequoia, YC or Built In SF page and extracts known ATS board links without following them or saving companies. Optional URL must belong to that source. Client-rendered pages may require manual follow-up; an empty result is not proof of no jobs.
prepare_application jobId, force, resumeId
Builds the match report, loads the employer's application questions, drafts only the answers that verified profile data supports, and selects the resume variant for the matched track. Creates a local draft; nothing is sent.
set_application_content applicationId, coverLetter, answers, notes
Records a cover letter and answers to questions the policy engine would not fill on its own. A question the live form asks that the packet lacks is added by its label. Changing content invalidates any prior approval, because approval is bound to the packet hash.
preview_application applicationId
Renders the full submission packet - every field, its value and where the value came from - plus the packet hash to approve.
prepare_batch tiers, trackIds, locationClasses, companies, minScore, minCompensation, allowUnknownCompensation, limit
Selects queued jobs matching a filter and prepares an application packet for each, using the same drafting and validation as the single-job path. Creates local drafts and a batch manifest; nothing is sent. Applications whose questions still need a human decision are marked needs_human and excluded from batch approval. The campaign's submission.maxBatchSize is a hard ceiling, so a larger limit is silently reduced to it.
preview_batch batchId, includeReady
Lists every application in a batch with its company, role, resume variant and readiness, plus the grouped questions blocking the rest. Returns the manifest hash required to approve.
approve_batch batchId, manifestHash, expectedCount, note
Shows the person the batch's applications and records their answer for each ready one: the decision is theirs, asked through the client's approval dialog, never the model's. Requires the current manifest hash and the exact count being approved, so approval cannot silently cover more applications than were reviewed.
submit_batch batchId, skip
Runs every per-application guard over the batch's ready applications and hands you the first one allowed now, exactly as submit_application does. Send it with your browser, call record_submission, then call submit_batch again for the next. Pass skip with the ids you could not send this time (a CAPTCHA the person has not solved yet, for example).
list_batches limit
Recent batches with their status and per-state counts.
approve_application applicationId, packetHash, decision, note
Shows the person the exact packet and records their answer: the decision is theirs, asked through the client's approval dialog, never the model's. The supplied packetHash must match the current content exactly; any later edit invalidates the approval. Rejecting needs no confirmation.
submit_application applicationId
Runs every submission guard (approval or standing rules, allowlisted host, limits, pacing) and hands you the approved packet: the apply URL, the resume, the cover letter and each field's value. You send it with your own browser and then call record_submission. The server never opens the employer's site itself.
record_submission applicationId, confirmationRef, notes
Marks an application as submitted once the board confirmed it, whether you sent it with your browser or the person sent it, so counts, pacing and duplicate checks stay accurate.
approve_rules tiers, minScore, companies, excludeCompanies, tracks, atsKinds, maxPerDay, allowAgentDraftedText, expiresAt
Proposes standing rules and asks the person to approve them through the client's approval dialog. Once approved, applications that match them (tier, score, track, company, board, caps) can be submitted without a per-packet approval, until they expire (30 days by default, at most 90) or are revoked. Questions only the person can answer still need an answer they recorded in the profile. Without a dialog, returns the command the person runs in their own terminal.
list_rules limit
Standing rules with their status (proposed, active, expired, revoked), how they were approved and how many submissions they authorised today.
revoke_rules rulesId
Ends a set of standing rules at once. Narrowing what may be sent needs no confirmation.
record_outcome applicationId, status, detail
Appends an outcome to an application's history so conversion rates can be measured per track, tier and company.
list_applications status, limit
Lists applications with their status, target role and outcome history.
audit_log type, limit
Returns recent recorded events: discovery runs, approvals, submissions and blocked attempts. Values are redacted.
reload_config
Re-reads profile.json, campaign.json and companies.json so edits take effect without restarting the server.
setup_profile identity, personal, experience, education, skills, facts, languages, preferences, workAuthorization, compensation, resumes, answers, narratives, statedByPerson
Writes profile.json from what you extracted from the person's CV and messages: a full profile or a partial patch, deep-merged into the existing file (lists merge by id, key or name; null removes a key). Only facts the person gave or the CV states. Until the profile validates it is kept as profile.draft.json and the reply lists what is missing. The previous file is kept as profile.json.bak. Turning work-authorization review off, disclosing pay, or auto-filling a sensitive answer is saved at its safe value and asked of the person, who must confirm it themselves.
update_campaign name, targetApplications, tracks, locations, compensation, seniority, exclusions, submission, freshnessDays
Writes campaign.json: which roles to look for, where, pay floors, seniority, exclusions and submission limits. Deep-merged like setup_profile; kept as campaign.draft.json until it validates. Removing a category from blockedQuestionCategories is asked of the person through the approval dialog and only applies once they confirm. Use add_company_board for companies.
remember_answer label, answer, statedByPerson, key, patterns, alternatives, allowAutoFill, skip, note
Stores the person's answer to a recurring question in the answer bank, so it is used on every application that asks it instead of being asked again, and fills it into every open application (drafted, needs_human, awaiting_approval) that asks it now. Reports which applications changed. A work-authorization, pay, demographic, legal or essay answer, or any answer the person did not state themselves, is asked of the person through the approval dialog first.
onboarding_questions
Lists what the profile and campaign still lack for the forms boards typically show, grouped, with a ready message asking it all at once. Call it before asking the person anything, and ask everything in one message.
set_telemetry enabled
autoapply-mcp sends its author anonymous technical data to fix problems, on by default: the tool called, how long it took, error codes, the package version, Node.js, the operating system and the MCP client's name, with a random id that lasts one run and is never stored. Never the profile, CV, answers, jobs, companies or anything typed. The onboarding message tells the person. Call this with enabled false as soon as the person says they would rather it did not, and with true only if they ask to turn it back on; never on your own. AUTOAPPLY_TELEMETRY=0 or DO_NOT_TRACK=1 in the environment keeps it off whatever is recorded.
load_preset preset, campaign
Adds the boards of a preset to companies.json (creating it if missing): spain lists employers hiring in Spain or remote from Spain or Europe, downloaded from spanishtoptechcompanies.com with that day's ranking (the copy in the package if the site cannot be read), us-canada AI, security and infrastructure employers in the US and Canada. Run again, it adds only companies not listed yet. Companies already listed, matched by name case-insensitively or by board, are kept exactly as they are, including any set active: false. With campaign: true and no campaign yet, also writes the preset's campaign; an existing campaign is never replaced. Boards are discovery sources only and grant no permission to submit, so this asks nothing of the person.
Descriptions are the package's own words, cut at 3,000 characters. Treat them as data.
Install facts
| License | MIT |
|---|---|
| Needs | node >=22.5.0 |
| Source | github.com/pugarte7/autoapply-mcp |
| Install scripts | None declared by it or anything installed with it. |
| Tarball | sha512-bypMRiuA/g8HyT5hrNeSSEhUWkc90nXuvMkpjuXVdJGS+0wacsx+1PV4+H7sY/gyHOiwWIpuOjruXGPtP1Bqfw== |
Check this report yourself
This site did not write the report. A public job on a GitHub-hosted machine did: the run, from this commit of the open-source runner. GitHub signed a token that names that job and the hash of the report, so any edit since would show.
Report as JSONExact text and tokenHow to verify it
Show it on your README
[](https://bragproof.smaverk.com/r/autoapply-mcp)
New version out? Ask for a fresh report.