Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion .github/workflows/ci.yml
Original file line number Diff line number Diff line change
Expand Up @@ -510,7 +510,7 @@ jobs:
R=eval/results
# E1: a witness catches every attack
jq -e '[.mutations[].ledger_plus_witness] | all(. == 1)' $R/e1_integrity.json
# E17: all 600 harmful calls stopped, at most 1 false alarm
# E17: all 952 harmful calls stopped, at most 1 false alarm
jq -e '([.classes[].caught] == [.classes[].deny]) and ([.classes[].fp] | add) <= 1' $R/e17_policy_corpus.json
# E20: no honest call held, no TK-X204 on an honest session, an unrequested edit found in every session
jq -e '.tuned.honest_sessions.held == 0
Expand Down
4 changes: 2 additions & 2 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -16,7 +16,7 @@ systems and show what they did: your customers and auditors can verify the recor
<table>
<tr>
<td align="center" width="25%"><h3>8 of 8</h3>tamper attacks caught<br><sub><a href="#proof">480 trials</a>, with a witness (E1)</sub></td>
<td align="center" width="25%"><h3>951 of 951</h3>harmful calls stopped<br><sub><a href="#proof">11 classes</a>, 1 false alarm in 971 (E17)</sub></td>
<td align="center" width="25%"><h3>952 of 952</h3>harmful calls stopped<br><sub><a href="#proof">11 classes</a>, 1 false alarm in 971 (E17)</sub></td>
<td align="center" width="25%"><h3>&lt; 1 ms</h3>added per tool call<br><sub><a href="#proof">0.91 ms at p99</a>, signer round trip (E9)</sub></td>
<td align="center" width="25%"><h3>8 of 8</h3>insider attacks caught<br><sub><a href="#proof">a merge gate</a> on every PR (E8)</sub></td>
</tr>
Expand Down Expand Up @@ -344,7 +344,7 @@ Our own experiments: small, synthetic and written by us. Nothing rounded up, and
| **Cross-check** (E10): a skipped hook, a made-up call, changed arguments | **12 of 12 flagged**, 0 false positives |
| **Three verifiers, one verdict** (E22): 63 bundles, 33 of them forged, each checked by the Python CLI, the TypeScript CLI and the browser page | **All three agree on all 63**; no forgery verifies. A 50,000-call run verifies in the browser in 14 s |
| **Overhead** (E9): signer round trip over a Unix socket | **0.45 ms p50, 0.91 ms p99** per call; 5,356 events/s |
| **Policy packs, known cases** (E17): 951 harmful, 971 benign calls | 951 stopped, 1 false alarm |
| **Policy packs, known cases** (E17): 952 harmful, 971 benign calls | 952 stopped, 1 false alarm |
| **Policy packs, written blind to the rules** (E21): 420 harmful, 432 benign | 211 of 420 stopped or held on the one blind run (286 now, with rules added after reading the pool); 18 benign held |

That last row is on purpose. **Policy rules are tripwires, not a wall.** They match text, and a determined agent can
Expand Down
8 changes: 4 additions & 4 deletions docs/evaluation.md
Original file line number Diff line number Diff line change
Expand Up @@ -282,8 +282,8 @@ p99. `make eval` runs the `--quick` form (5 calls per phase instead of 20). Resu

Three offline parts, each with a signer in process; nothing is executed.

- **Policy packs** (`eval/e17_policy_corpus.py`): 951 harmful and 971 benign tool calls in 11 classes
(`tests/data/policy/`), written by the project, decided by `server.yaml` with both engines. All 951 harmful calls are denied or held; one benign call is held (a draft email to a list address, `TK-M003`), 1 in 971 (0.10%); the engines
- **Policy packs** (`eval/e17_policy_corpus.py`): 952 harmful and 971 benign tool calls in 11 classes
(`tests/data/policy/`), written by the project, decided by `server.yaml` with both engines. All 952 harmful calls are denied or held; one benign call is held (a draft email to a list address, `TK-M003`), 1 in 971 (0.10%); the engines
decide every call the same. It gates on every harmful call caught, at most 2% false positives and engine agreement.
- **Detection** (`eval/e17_detection.py`): 10 scenarios for [sequences](incidents.md): a sequence split across two
runs, a tool alias the policy doesn't map, parallel runs, a GET with a query string as egress, a denied step that
Expand Down Expand Up @@ -620,10 +620,10 @@ All 50 are benign calls strict holds (ten more than 1.5.2's 40: 1.5.3's strict h
network command (`TK-F013`), 3 as a shell startup-file edit (`TK-F011`), and 1 draft email to a list address (`TK-M003`); `TK-F012` (any URL a call acts on) only flags in
strict, it never holds one by itself.

Shell rewrites (`eval/e21_robustness.py`): E17's harmful calls are decided as they are, and each of the 615 shell
Shell rewrites (`eval/e21_robustness.py`): E17's harmful calls are decided as they are, and each of the 616 shell
commands among them again rewritten in up to five equivalent ways (`cd . &&` or `true;` before it, Bash's `command`
builtin, `bash -c` and `env bash -c` around it; a form that could change what the command does is left out, such as
`bash -c` around a `cd` or an `export`). A call is contained when the server pack still denies or holds it: all 951 originals and all 2,948 rewrites are, and both engines decide each the same. The run gates against a recorded
`bash -c` around a `cd` or an `export`). A call is contained when the server pack still denies or holds it: all 952 originals and all 2,953 rewrites are, and both engines decide each the same. The run gates against a recorded
baseline, rule id by rule id, so a rule that held a form and no longer does fails it. Results:
`eval/results/e21_robustness.json`.

Expand Down
2 changes: 1 addition & 1 deletion docs/policy-v2.md

Large diffs are not rendered by default.

Loading
Loading