Skip to content

Candidate rights and appeals

Fairness to candidates is both a legal requirement and a product principle. This page documents what a candidate is told, what they can see, and how they can contest a result — including the right, under GDPR Article 22, not to be subject to a solely automated decision with legal or similarly significant effects without safeguards.

Employer and SkillFoundry share this responsibility

SkillFoundry provides transparency tooling and a human-review path in the product. The hiring decision, the appeal outcome, and any accommodation are the employer's to grant. Where SkillFoundry is the controller (self-serve candidates), SkillFoundry handles requests directly.

What candidates are told

  • Before an assessment: the browser IDE-mode setup discloses telemetry before the candidate opts in, and the in-editor AI-policy banner states the integrity rules in force. See Monitoring and data collection.
  • AEDT notice (employer-provided): where the tool is an AEDT, the employer must give candidates advance notice (≥10 business days under NYC LL144) of the use, the characteristics assessed, and how to request an alternative process. SkillFoundry supplies reusable "characteristics assessed" language — see Bias audits and adverse impact.

What candidates can see about themselves

  • Live telemetry during the session. Candidates can view the complete event stream as it is collected — the same data reviewers' scores are built from. In VS Code, the Assessment Panel's Behavioral Data tab refreshes every ~2 seconds and the command SkillFoundry: View Collected Events (Transparency) shows the full summary.
  • Their behavioral report. Where the organization enables it, candidates see their own report — overall score, per-skill breakdown, strengths, and improvement recommendations.
  • Their data, on request. An access request returns a structured export of everything held about them: profile, consent records, audit history, tasks, submissions (including code and feedback), conversations, and organization membership. See Data retention and erasure.

This "what you see is what the platform has" design is deliberate: it removes the information asymmetry that makes opaque behavioral scores objectionable to regulators.

Exercising rights

A candidate can request to access, correct, delete, port, restrict, or object to processing of their data. Practically:

  • Employer-run assessment: the candidate contacts the employer (controller); SkillFoundry assists the employer and forwards any request a candidate sends it directly.
  • Self-serve candidate: the candidate contacts SkillFoundry support / privacy@skillfoundry.dev.

Each request is tracked as a data-subject request with a status and is audited end-to-end. See Data retention and erasure.

The right to human review and appeal

SkillFoundry is built so that automated output does not, by itself, reject a candidate:

  • Integrity flags trigger human review — never auto-rejection. Where AI-usage detection is active, a flag routes the submission to a human reviewer; the flag is a prompt to investigate, not a verdict. See Reviewing candidates.
  • Reviewers can override the automated score. Manual review applies a points adjustment with a required reason and optional candidate feedback, recorded with the reviewer's identity — so an appeal can be actioned and evidenced.
  • Working-process results are context, not filters, and reviewer guidance says so explicitly.

To satisfy Art. 22 safeguards and good practice, we recommend employers offer, and SkillFoundry's controls support:

  1. Notice of the outcome and that automated tooling was used.
  2. A channel to contest the result and request human review.
  3. A qualified human re-review of the submission (code diff, notes, transcript, and the supporting evidence — not just the number), using the manual-adjustment mechanism to correct the record where warranted.
  4. A written explanation referencing the specific parts of the result, since scores are broken down with supporting evidence rather than a black box.
  5. An alternative-process / accommodation path for candidates who request one (see below).

Accessibility and accommodations

Because working-style measures can correlate with disability or neurodivergence (see Bias audits and adverse impact), employers should offer accommodations such as extended time, an alternative assessment format, or excluding time-based measures from the evaluation. SkillFoundry does not use webcam, microphone, or screen recording, which avoids a class of proctoring accommodations, but employers remain responsible for reasonable-adjustment obligations under the ADA, the Equality Act, and equivalents.

What we tell candidates we never do

Restating the guarantees from Monitoring and data collection, because candidate trust depends on them:

  • No keystroke logging, ever.
  • No clipboard contents.
  • No webcam, microphone, or screen recording.
  • No collection outside the assessment (and, in VS Code, outside the assessment workspace folder).
  • No collection before the assessment starts or after it is submitted.

Roadmap

  • A self-service candidate privacy center (view consent, download data, submit rights requests) backed by the existing privacy-dashboard payload.
  • An in-product appeal/contest button on candidate-visible reports that opens a tracked review request.