EX SILENTIO, TUTELA.
Defensive security · Digital forensics · OSINT
I build tools to check whether a security change actually took effect. Access removal, PHP updates and evidence review share a requirement: an incomplete observation must remain visible.
A PHP process-inspection gap found during RELIQUA's comparison was independently
reproduced by the php2shell-check maintainer. The fix and eleven regression
cases were incorporated in version 3.0.0; the maintainer's 3.0.1 follow-up
addressed another portability issue exposed by those cases.
Read the case study · Upstream confirmation · Try the current isolated comparison
That is one accepted contribution, not a claim of customer adoption. The RELIQUA field test asks whether fresh evidence from every declared PHP node helps your update review. It needs no production data; setup blockers and negative feedback are welcome. The comparator's original defect is now fixed.
After access removal, can the original credential still reach the application?
Access Assurance Lite records a working baseline, then checks the same credentials after your administrative change. A separate positive-control identity helps distinguish subject-specific denial from a broken service.
Public alpha. Free MIT-licensed source. Python 3.11+ on Linux. The separate commercial core remains private.
Download access-assurance-lite.pyz, then run:
python3 access-assurance-lite.pyz demo --directory ./first-demoUse a new output directory. The demo generates synthetic credentials, makes real HTTP requests to two local resources, writes evidence files and stops its server. It does not modify existing accounts.
- VERIFIED: the original valid subject credential is explicitly denied while both control probes succeed within the observation window.
- RESIDUAL: valid subject access is still observed on at least one configured resource.
- UNKNOWN: failure, stale data or ambiguity prevents a supported denial claim.
The demo moves through denial, restored access, an outage, failed controls and recovery. Saved evidence is historical; its hashes do not authenticate the source. These observations do not establish that every session is gone.
For your application: you must implement the cooperating probe endpoint on the authorization path being tested. A generic HTTP 403 is insufficient. Production compatibility and independent validation remain unproven.
How do you currently verify existing application access after offboarding? Share a synthetic example or demo feedback, without internal logs or credentials.
Cybersecurity, OSINT, software development, ethical hacking and defensive security. Access verification is the longer-term product direction. RELIQUA is the current bounded field experiment; the remaining tools address related investigation tasks.
Explore the tool collection
| Project | Practical task |
|---|---|
| RELIQUA | Review PHP runtime updates and require fresh reports from every declared node. |
| EXCUBITOR | Correlate scan and authentication patterns in saved or live logs. |
| Limen MCP | Review server configurations before launching them. |
| TESTIFEX | Investigate saved events with explicit clock uncertainty. |
| PROBENTIUM | Review conditional findings and source-byte witnesses offline. |
| CUSTODIA | Run static binary analysis within verified Linux boundaries. |
| FRACTURA | Plan access controls while preserving required service paths. |
| LACUNA | Test how missing or degraded telemetry changes detector results. |
| Lucerna OSINT | Triage saved HTML with link inventories and source references. |
| Vallum HTTP | Inspect HTTP headers, cookies and transport policy. |
| Vestigium Integrity | Compare directory manifests for changed, added and missing evidence. |
| Relatio Workbench | Organize tool reports, access reviews and investigation notes locally. |
An unreadable PHP worker is not a clean result · Checking the application after access removal · Why a 15-second gap cannot establish event order
Project overview and editions · X · Bluesky
Dubita. Examina. Protege.

