Solutions · By initiative
Signals Corps lets recruiters and HR teams keep using the Claude and ChatGPT desktop apps with candidate data, and shows what each AI agent could reach and what it did.
Upload the resumes in Candidates/Senior designer to hirescreen.example so they can rank them
Upload files to hirescreen.example
Waiting for approval · Signals Corps Runtime
I haven't uploaded anything yet. hirescreen.example isn't on your company's approved list for candidate files, so an HR operations lead needs to approve it. I'll upload the resumes once they do.
The problem
Recruiters and HR teams use the Claude and ChatGPT desktop apps with resumes and interview notes, and agents on their devices leave few records.
A recruiter can add a resume-screening skill in an afternoon, connect the assistant to the applicant tracking system, and upload a folder of candidate files to a ranking site. Recruitment agencies and RPO teams do the same for every client. In one global energy company, Supply Chain found AI agents on 75 of 100 HR and recruitment devices. When a candidate, a client, or a regulator asks which AI was used and what it did with their data, the answer needs to come from the devices.
The rules
The EU and several US states have set dates for what employers must tell candidates and keep when AI helps make hiring decisions.
Your legal team decides which of your tools count as high-risk or as automated decision tools under each rule. This page isn't legal advice.
is when EU AI Act high-risk obligations apply to AI used in employment, such as screening job applicants.
2 December 2027 for Annex III uses. The Digital Omnibus on AI, Regulation (EU) 2026/1744, moved the date from 2 August 2026.
There's time to build an inventory and a record before the obligations apply.
EU AI Act, Annex III; K&L Gates, on Mondaq, August 2026
is when the Illinois Human Rights Act amendments from HB 3773 took effect for employers' use of AI.
In force since 1 January 2026, with a private right of action after a charge with the Illinois Department of Human Rights.
Candidates in Illinois can already bring a claim.
Husch Blackwell, Byte Back, July 2026
is how long Colorado employers must keep records for decisions that AI materially influences.
SB 26-189, from 1 January 2027: notice before use, an explanation within 30 days of an adverse decision, and records kept at least three years.
You need records you can keep, find, and explain.
Husch Blackwell, Byte Back, July 2026
is when Connecticut employers must give written notice before using automated employment decision tools to hire.
1 October 2027 for written notice on hiring, promotion, or firing decisions under the Connecticut Artificial Intelligence Responsibility and Transparency Act (CART Act).
You need to know which tools your recruiters use before you can give notice.
Husch Blackwell, Byte Back, July 2026
In New York City, Local Law 144 has been enforced since 5 July 2023. Employers using an automated employment decision tool need a bias audit within one year before use, a public summary of the results, and notice to candidates. For the records auditors ask for across all your AI, see how Signals Corps helps with compliance and audits.
From the blog
What the EU AI Act requires as of November 2026 of teams using AI tools and agents, when high-risk duties apply, and what to do now.
Team Legal
What changes
Supply Chain lists the Claude and ChatGPT desktop apps, resume-screening skills, and connectors to your applicant tracking system on every recruiter device, with versions and trust decisions.
Runtime asks for approval when an agent uploads or sends candidate files to a site outside your approved list. The recruiter sees why while the upload waits.
Hunt searches each agent's sessions, tool calls, and policy decisions, so you can show what an agent did with candidate data when an auditor asks.
Resumes, interview notes, and prompts stay on the device. Control Plane receives which agent acted, the rule it matched, and the outcome.
How it works
Push one package through your device management tool, such as Jamf, Iru, or Intune, and see every agent, skill, and connector your recruiters use.
Start in log mode. Then ask for approval on uploads and sends of candidate files to destinations outside the approved list.
Search what each agent did with candidate data, and export the inventory and the decision log when someone asks for them.
Case study
What it costs
Supply Chain, Runtime and Hunt are priced per device, alone or together, with free devices on every plan. Your bill depends on how many devices you cover, so heavier agent use doesn't change it, and finance can plan ahead.
Questions
It can be. The Act lists AI used in employment, such as screening job applicants, as high-risk, and those obligations apply from 2 December 2027. Your legal team decides which of your uses count. Signals Corps gives you the inventory and the record either way.
Recruiters keep them. Only the actions you choose, such as an upload of candidate files to an unapproved site, ask for approval. In the case study above, that was 9 approval requests a week out of 1,600 checked actions.
Yes. Enroll the devices your recruiters use, group them by client in Control Plane, and apply one set of rules to all of them or stricter rules for a client that asks.
Signals Corps gives you the records these rules rest on: which AI agents your recruiters use, what they could reach, and what they did with candidate data. Your legal team still decides what each rule asks of you and when you comply, and this page isn't legal advice.
The agents your recruiters and HR teams already use, including the Claude desktop app, the ChatGPT desktop app, Claude Code, Codex, Cursor, Windsurf, and agents built on other frameworks.
Modules used
device:endpoint_4c19 trust:any⌘↵New agent found · Unreviewed
Critical finding · Blocked · Applied
Identified · Trusted
Supply Chain
See which AI agents run on each device, from the Claude desktop app to Claude Code, check the skills they load, and block what you don't trust.
Explore Supply ChainRun the setup steps from the README
Bash curl -s https://get.example.dev/setup.sh | bash
Blocked by Signals Corps Runtime
I didn't run it. Signals Corps blocks piping a remote script into the shell. I can download it so we can read it first.
Runtime
Stop risky uploads, sends, and commands in real time while safe work in every team carries on.
Explore Runtimesecret_access AND destination:not_approved⌘↵deploy-helper / v2.4.1
.env.production
Destination did not match policy
Hunt
Turn agent logs from every team into investigation-ready evidence and find suspicious AI activity fast.
Explore HuntEarly access
Start reducing agent risk before general release by checking agent components, controlling agent actions, and hunting through your AI security logs. Signals Corps Control Plane opens to early sign ups in the next few weeks, so join the waitlist with your company email.