Prompt2PWN has three parts in one Intigriti program. The asset groups and In scope sections use the same names: 1. Prop Bounties: Specific challenges with defined finish lines and one winner per challenge, including Riviera ULXC isolation escape and Shadow IT to protected internal access. Each offers up to $10,000 under its published terms. These challenges remain conditional until the exact targets, starting access and payout schedules are published. More Props may be added during the event. 2. Smarter Dropbox: First access to new features and endpoints in Dropbox's latest product release, building on Dash with more integrated AI assistance, including Dropbox CoWork. Use your AI hackbot to investigate protected-content access, permission boundaries, privilege escalation and unauthorized AI-driven actions. The testing objectives guide research; qualifying unique findings follow the Progressive Bounties policy, up to $5,000 per finding. 3. Telemetry Challenge: Rewards for substantive recon data and tool outputs, even when no vulnerability is found. One $100 or $250 base award per researcher, plus an equal share of the unlocked collective bonus pool, subject to the first-50-contributor cap and package requirements. Telemetry submissions remain open through Sunday, November 15, 2026. See each matching In scope section for authorized targets, completion requirements and submission instructions. Telemetry rewards are separate from vulnerability and Prop awards. The Reward policy governs validation, duplicates, upgrades and overlapping payments. Use approved organizer-provided test accounts when issued, or create Free Standard Team event accounts through the published signup flow using your Intigriti alias. The 30-day free trial remains an option. Cancel trial accounts within 30 days to avoid a credit-card charge. Contact Dropbox support via Discord for access or unexpected purchase issues. You need an Intigriti account to participate. Invitations and submissions can proceed while ID verification is being completed; verification is required before payment. The program is currently suspended, so submissions cannot be sent until it reopens.
Progressive Bounties
Awards are total per unique finding. Assessment considers verified impact, likelihood, prerequisites, victim interaction, sensitivity, reach and the permission boundary crossed.
| Qualifying outcome | Approved total per finding |
|---|---|
| Entry: restricted non-public file metadata or account/user/tenant identifiers disclosed across a permission boundary, with evidence of expected denial. | No bounty |
| Low: protected ownership, membership, sharing or account-to-resource relationships; or a usable unauthorized capability with protected context beyond a basic identifier. | $100 - $250 |
| Medium: bounded confidential-field disclosure or protected-record modification matching a published outcome, with meaningful impact. Sensitive, reproducible outcomes may warrant the ceiling. | $250–$1,000 |
| High: effective standard-user takeover, material protected-content access or consequential unauthorized action. Prerequisites and constrained reach affect the total. | $1,000–$2,500 |
| Critical / Exceptional: substantial administrative authority, broad arbitrary protected-content access, multiple-identity control or new service compromise with protected authority. | Up to $5,000 |
| Public/intended information or identifier capabilities; ordinary enumeration; guessed IDs without disclosure; unverified scanner output; hypothetical chains. | No vulnerability award |
Qualifying Medium findings start at $250; High findings start at $1,000. Upper amounts are ceilings, supported by stronger impact and a documented rationale. Native severity bands are an administrative display; demonstrated impact and these terms determine the approved total.
Entry findings qualify independently. A private document ID disclosed to unauthorized account A may earn $100; added protected ownership/sharing context may justify $250. You may qualify directly at any tier. Additional IDs, prompts or endpoints alone do not increase the award. A snippet, limited admin action, cross-tenant label or lab marker does not establish severity by itself.
Progression and payment
- Submit promptly. The earliest complete submission later validated has priority; placeholders reserve none.
- Upgrade the same report. Add verified impact, starting permissions, minimal reproduction and the requested revised total. Report independent flaws separately and link them.
- Triage approves the revised impact and total. A severity change alone does not trigger payment.
- Additional payment = revised approved total − prior vulnerability/upgrade payments for that finding. An unchanged total receives no top-up.
Duplicates, chains and other rewards
- Same root cause and remediation = one finding. New prompts, accounts, endpoints or outputs alone do not earn separate awards.
- Independent flaws are assessed separately; overlapping impact is counted once. Chains do not make every component Critical. Known-issue replays receive no new-finding award. Seeded challenges require explicit terms; distinct flaws and mitigation bypasses need verified impact.
- Telemetry is paid separately and is not deducted from vulnerability upgrades. A $100 finding plus accepted $250 telemetry pays $350 before any pool share. A finding alone does not qualify for the telemetry pool.
- Enabled Props replace overlapping vulnerability awards: deduct prior overlapping payments from the actual approved Prop total. Retests/fix validation require separate terms. Upgrades follow the published deadline; testing after the window requires renewed authorization.
Telemetry bounty and collective bonus
$100 or $250 per researcher, independent of vulnerability rewards. Qualifying packages capture substantive bot testing, including failed or blocked runs.
Telemetry submission guide and starter kit
| Tier | Base reward / researcher | Acceptance criteria |
|---|---|---|
| Useful contribution | $100 | Attributable real testing: handle/run ID, bot/setup, starting access, timestamped actions/results, summary, human assistance and limitations. Sanitized core package or native logs with field mapping. |
| High-quality contribution | $250 | All base requirements plus a reconstructable action/result/evidence sequence, retries/branches, changes in approach, model/script/human attribution, available correlation and support for claimed effects. |
One highest base award per researcher/event. Additional runs strengthen the same contribution; a $100 → $250 upgrade pays $150 more. Volume, copied logs, unchanged templates, fabricated activity, synthetic examples presented as real runs, and unsafe testing do not qualify.
50 rewarded places. Allocation follows Intigriti receipt time of each unique researcher’s first complete qualifying package, not review order. Empty, duplicate or nonqualifying packages reserve no place; corrections and additional runs use the same place. Organizers validate eligibility and close rewarded intake at 50. Beyond the cap, no telemetry base/pool payment applies; vulnerability eligibility is independent.
Collective bonus: each milestone adds to one cumulative pool, split equally among all admitted contributors on top of base and vulnerability awards.
| Qualifying researchers | Added at milestone | Total pool | Share at that exact count |
|---|---|---|---|
| 10 | $1,000 | $1,000 | $100 each |
| 20 | $2,000 | $3,000 | $150 each |
| 30 | $3,000 | $6,000 | $200 each |
| 40 | $4,000 | $10,000 | $250 each |
| 50 — cap | $5,000 | $15,000 | $300 each |
Between milestones, all contributors share the last unlocked pool: 15 share $1,000; 25 share $3,000. Fewer than 10 unlocks no pool. Both quality tiers count equally. Count researchers with attributable substantive contributions, not accounts/runs/bots. A shared dataset does not automatically qualify every team member.
Final count follows the published submission/correction deadline and review. Round equal shares down to cents; distribute remaining cents in qualifying-package receipt order. At 50, total telemetry payout is $400 or $550 per researcher, including the $300 pool share.
Submission package: hackbot-telemetry-<run_id>.zip with the following checklist, matching the linked participant guide:
- Required:
manifest.jsonandsummary.mdfor researcher/run/setup, actual activity, review and limitations. - Required activity records:
actions.jsonl, or mapped native JSON, JSONL or CSV logs. - Required capture setup: sanitized
capture-setup.mddocumenting configured capture and pre-flight proof. - Required session lifecycle:
sessions.jsonl, or equivalent mapped native lifecycle records. - Required for native logs or another schema:
field-mapping.md. Keep each source's actual schema label; do not relabel native logs as schema 1.1.
Capture actual activity, UTC timestamps, available correlation, model/setup, permissions, human involvement and gaps/unknowns. Paper covers optional decision/interaction records, inventory and minimal evidence. Hidden reasoning and proprietary prompts are never required; visible AI messages are optional.
Submit a human-reviewed, sanitized package privately through Intigriti. Attach/reference it in a report without a finding, use the announced telemetry-only route and label Telemetry only — no vulnerability, following its classification/payment terms.
Conditional Prop Bounties
Riviera ULXC and Shadow IT each offer up to $10,000, once, to the earliest complete validated submission. The ceiling depends on protected authority and practical impact; a lab marker alone does not establish production-scale compromise. Prior overlapping vulnerability payments are credited against the approved total. Independently qualifying telemetry remains payable within its cap.
Props remain disabled until exact assets, starting access, finish line and a separately approved payout schedule are published. See In scope for the objectives. The optional agent VM escape is unpriced and disabled.
For obvious reasons we can only allow submissions or applications for our program with a valid Intigriti account.
It will only take 2 minutes to create a new one or even less to log in with an existing account, so don't hesitate and let's get started. We would be thrilled to have you as part of our community.