The scope and responsibilities that apply when using GudTek.
Last updated 27 October 2026
1. What the service provides
GudTek provides a holder-gated agent launcher, launch-planning tools, and direct read-only token or wallet inspection. Agents perform bounded analysis of selected public Solana data, public GitHub source, or repository search metadata. Results are informational observations produced from available evidence and explicit rules. The public protocol explorer offers reference explanations and allowlisted repository snapshots. The holder network displays bounded public token-balance observations and threshold comparisons; its ranking is by observed holdings, not reputation or trading performance. Features, coverage, and availability can change as the application develops.
The current service is not an independent audit, certification, insurance product, managed custody service, investment adviser, exchange, or trading venue. It does not guarantee that a token, wallet, repository, project, or launch is safe.
2. Account, holder access, and saved records
You are responsible for access to your sign-in method, connected wallet, device, and browser. Use only accounts and credentials you are entitled to use. Do not place secrets or private-key material in agent names, targets, research keywords, or project fields.
Creating an agent profile requires an authenticated account. Executing a live agent run additionally requires the configured token holding in verified wallets linked to that account. Holder eligibility is checked on the server for each run. Access remains closed when the mint, positive minimum, or verification services are unconfigured, and may be denied when a check fails or holdings no longer meet the requirement. An earlier successful check is not permanent access or a commitment about the token’s value.
Project stages and checklist completion are statements made by the user. Marking a project “Team ready” does not represent endorsement or approval by GudTek. You remain responsible for the accuracy of project information, selected agent targets, and evidence behind each completed task.
3. Authorized public-source use
Use the service lawfully and without interfering with other users or its infrastructure. Do not bypass authentication, holder checks, or rate limits; attempt to access another user’s records; submit malicious content; overload the service; or misrepresent a result as an independent certification.
Repository analysis is limited to public GitHub material you are authorized to access and analyze. Public availability does not remove applicable licenses, attribution requirements, or other rights. Do not use the service to retrieve private repositories, expose credentials, or circumvent a source’s access controls.
Public-address and source inspection do not authorize access to private systems, credential use, exploitation, or active testing against a website or network. The current workers do not run target code or active exploits. Any separate testing needs appropriate authorization.
4. Interpreting results
A finding reflects a specific observation at the reported time or source commit. A passing rule is limited to that rule; an unknown status means no supported conclusion was reached. The token configuration score is a heuristic over implemented automated checks, not a probability of safety or a promise about future performance.
Repo Recon samples selected files and reports source-pattern observations. It does not prove an exploitable vulnerability, repository originality, contributor authenticity, whole-program correctness, or deployment equivalence. Tech Discovery supplies metadata leads, not endorsements or verified innovation. A completed run can have partial evidence; failed or cancelled execution is not a passing review.
Results do not establish liquidity safety, sellability, beneficial ownership, team identity, key custody, or investment suitability. Provider errors, incomplete data, unsupported behavior, and later changes can affect results. Review the methodology and coverage limits before relying on a report.
5. Decisions and transactions
You decide whether and how to act on findings. Independently verify addresses, instructions, code, licenses, authority changes, and transaction effects before taking action.
Agent and scan workflows do not sign transactions, approve token spending, revoke permissions, transfer assets, install repository dependencies, or modify source code. A connected wallet supports authentication and read-only holder eligibility checks. No agent automatically protects a project, repairs a finding, or recovers funds. A public address or repository appearing in the application does not imply its owner’s endorsement.
6. Results and attribution
When sharing plans or results, preserve the target, source links, network or commit, timestamp, example-data labels, and coverage limitations so recipients can understand the scope. Do not remove qualifications to imply broader verification.
Only upload artwork you are entitled to publish. Artwork is stored at a public URL. If you choose to publish an agent, its target and latest run evidence are visible through the public directory. Publishing is not a claim that its code, coin, or findings have been independently verified.
Third-party names, marks, links, and public-source references identify their respective services or sources. They do not imply a partnership, endorsement, certification, or affiliation. The isolated demo uses synthetic fixtures; its agents, projects, logs, and results must not be presented as real reviews.
7. Execution, availability, and saved data
Each agent launch starts one bounded background run. It is not a recurring task, durable monitoring service, or promise of completion. RPC, GitHub, authentication, hosting, and database providers can be unavailable or change behavior. Time limits, provider failures, and host interruptions can stop a run; retries require a new launch.
Cancellation is recorded immediately and stops further work at the next checkpoint; an issued request may finish first. A cancelled run saves no final result. The application does not promise continuous monitoring, automatic rescans, incident detection, alerts, or uninterrupted availability.
Account agent configurations and run history are stored until removed through available controls. Delete an inactive agent to remove its associated history. Guest projects and direct scan history depend on browser storage; the demo uses separate session storage. Keep copies of records you need. See Privacy & data use for storage and deletion details.
8. Changes and applicable rights
The application may update these terms as features and operations change. The date above identifies this version. Review material changes before continuing to use new or changed features.
Nothing on this page is intended to exclude rights or remedies that cannot lawfully be excluded. No particular governing jurisdiction, arbitration requirement, paid subscription, refund policy, or service-level commitment is created by this notice.