Tenex research search log
Research date: 2026-10-09. Target domain: tenex.co.
Scope
Explain the company’s service, discovery method, agent design, software delivery system, controls, and evidence of results. Find useful company and founder articles in addition to the supplied website. Preserve the distinction between current services, internal tools, proposed methods, and guest case studies.
Queries and paths
The following query text preserves the research record.
| Query or discovery path | Result |
|---|---|
site:tenex.co blog Tenex |
Company blog and playbook library |
Tenex tenex.co founders technical writing |
Company identity and unrelated companies with the same name |
"Tenex" "Alex Lieberman" "Arman" interview engineering story points |
Original Latent Space interview publication |
"Tenex" "MetaHarness" github |
No public MetaHarness implementation found |
"Arman Hezarkhani" blog Tenex |
Founder references and LinkedIn articles; LinkedIn articles were not fully reviewed |
"Tenex" "Alex Lieberman" customer case study |
Founder anecdotes and company references; no controlled customer return study found |
site:latent.space "Lieberman" "Tenex" |
Original publisher account, S21 |
site:vercel.com "Tenex" |
No separate partner confirmation found; founder announcement saved as S28 |
site:anthropic.com "Tenex" |
Older search excerpt; current page redirects and lacks the claimed mention |
site:github.com "Tenex" "MetaHarness" |
No public implementation found; absence remains a search limit |
| Company service, blog, playbook, and careers links | S01–S18 and S22–S24 |
| Founder website and article links | S19–S20 and S25–S29 |
| Employee skill article link to GitHub | Public repository, skill, and trace reference, S31–S33 |
| Founder article list | Recent agent sprint guide, S34 |
The archive preserves source URLs and outbound links in the manifest. Queries establish discovery scope. They do not prove that no other evidence exists.
Access and attribution decisions
- Exclude tenex.ai cybersecurity material and unrelated Tenex investment firms.
- Treat HTTP 200 job shells as partial captures.
- Treat the podcast page as a written publisher account, not a complete spoken transcript.
- Keep guest methods and results separate from Tenex’s own deployments.
- Check a current page after a search excerpt names a partner.
- Preserve the current redirect instead of the older excerpt as primary evidence.
- Inspect public skill text without installation or execution.
- Keep prototype descriptions separate from production capabilities.
Public HTTP capture recovered the founder sprint article and raw skill files after the web reader failed to fetch them. No account registration, newsletter subscription, paywall bypass, or private source access formed part of this research.
Remaining gaps
No controlled Tenex throughput test, audited financial result, or customer-authored return study was found in this search. MetaHarness executable code, the full skill registry, and the custom agent runtime remain undisclosed in the reviewed sources. The public skill establishes an instruction artifact. It does not establish runtime accuracy or security. The report states the questions that could change these conclusions.