Slaine

4 ms·

1. 1.▲
   by Slaine 8mo ago

   Yeah, totally agree. Integrity mostly answers the “what happened” part. The idea is that once the sequence of events is provably intact, you can attach the decision context to it — things like policy snapshots, inputs/prompts (or hashe
2. 2.▲
   Show HN: I built a tamper-evident evidence system for AI agents(guardianreplay.pages.dev)

   2 points by Slaine 8mo ago|2 comments
3. 3.▲
   Show HN: Deterministic NDJSON audit logs – v1.2 update (structural gaps)(github.com)

   1 points by Slaine 8mo ago|0 comments
4. 4.▲
   by Slaine 9mo ago

   To be clear: this repo is proof-only. The idea was to see whether the artifacts themselves (two NDJSON logs) are convincing without explanation. If they’re not, I’d really like to know why.
5. 5.▲
   Show HN: Deterministic event logs with explicit gap markers (NDJSON proof)(github.com)

   1 points by Slaine 9mo ago|1 comments
6. 6.▲
   by Slaine 9mo ago

   Small clarification: the demo is intentionally boring and minimal. The point isn’t the UI — it’s the exported trace. I’ve been running the same test multiple times with the same seed and network profile and diffing the JSON. The timings mov
7. 7.▲
   Show HN: A small reference demo for separating UI stalls from network delays(deterministic-stream-demo1.pages.dev)

   1 points by Slaine 9mo ago|1 comments