gitea-admin/CASAN
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat(plan-01): Phase 1 — relocate harness code to packages/casan-harness (symlink facade)

Browse Source 
Physically move the pure-code subtrees out of .specify into the package, leaving
compat symlinks at the old .specify/<dir> paths so every existing reference (internal
CASAN_HARNESS_ROOT + external CI/docker/mjs) keeps resolving. Runtime state stays put.

Moved (git mv): scripts/ tests/ security/ templates/ config/ governance/ memory/
  .specify/<dir>  ->  packages/casan-harness/<dir>   (+ .specify/<dir> symlink)
Stays in .specify (state/governance/domain, handled later): logs/ agentops/ level5/
  init-options.json traceability-map.json

Python `.resolve()` self-location followed the compat symlink into packages and lost
the app root; generate-casan-demo-context.py, generate-agentops-dashboard.py and
dashboard-server.py now walk UP for the `.specify` state marker instead of a fixed
parent depth (fixes "missing trace files" in run-casan4).

Full gate: PASS=64 FAIL=0 SKIP=3 (CASAN_CI_STEP_TIMEOUT_SEC=1200 — track-a ~450s runs
close to the 600s default and can tip over under load; this is timing variance, not a
regression — it passed cleanly with headroom). Runtime log/audit artifacts kept unstaged.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This commit is contained in:
thanhnv
2026-07-08 00:06:00 +09:00
co-authored by Claude Opus 4.8
parent 2c765c9a45
commit 664bd1f00c
229 changed files with 268 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files
AINative_OKR_CASAN5/packages/casan-harness/tests/phase-sec14-tests.sh
+72
View File
@@ -0,0 +1,72 @@
#!/usr/bin/env bash
set -uo pipefail
# CASAN Plan-16 SEC-14 (M-03 / SC-03) — model-digest env-override is NOT trusted
# in enforced mode.
#
# current_digest() honored CASAN_MODEL_DIGEST unconditionally, so an attacker who
# swapped the local model could also set CASAN_MODEL_DIGEST to the pinned value
# and pass verification. Under CASAN_PROFILE=prod (or CASAN_MODEL_DIGEST_STRICT=1)
# the override is ignored and the digest must come from the live model backend.
# This proves:
# * dev/CI: override still honored (pin + verify deterministic) — regression,
# * dev/CI: a swapped override is a MISMATCH (control still works),
# * prod: a forged override equal to the pin is NOT accepted (fail-closed),
# * strict flag: same enforced behavior without a full prod profile.
#
# Deterministic; hermetic (temp pin file; live backend pointed at a dead port).
SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
source "$SCRIPT_DIR/../scripts/bash/casan-paths.sh"
PROJECT_ROOT="$(cd "$SCRIPT_DIR/../.." && pwd)"
MDC="$CASAN_HARNESS_ROOT/scripts/bash/model-digest-check.sh"
WORK="$(mktemp -d)"
trap 'rm -rf "$WORK"' EXIT
PIN="$WORK/model-digest.pin"
DEAD_OLLAMA="127.0.0.1:1" # unreachable → live digest fetch fails deterministically
PASS=0; FAIL=0
pass() { echo "PASS: $1"; PASS=$((PASS + 1)); }
fail() { echo "FAIL: $1"; FAIL=$((FAIL + 1)); }
echo "===== Plan-16 SEC-14: model-digest env-override not trusted in enforced mode ====="
# Setup: pin an approved digest in dev mode (override honored).
CASAN_MODEL_DIGEST_PIN="$PIN" CASAN_MODEL_DIGEST="digest-approved" \
bash "$MDC" pin ornith:9b >/dev/null 2>&1
# 1) dev/CI regression: matching override verifies OK.
rc=$(set +e; CASAN_MODEL_DIGEST_PIN="$PIN" CASAN_MODEL_DIGEST="digest-approved" \
bash "$MDC" verify ornith:9b >/dev/null 2>&1; echo $?)
[[ "$rc" -eq 0 ]] && pass "dev: matching override verifies OK (regression)" \
|| fail "dev: matching override should verify OK (rc=$rc)"
# 2) dev/CI regression: swapped override is a MISMATCH (block).
rc=$(set +e; CASAN_MODEL_DIGEST_PIN="$PIN" CASAN_MODEL_DIGEST="digest-swapped" \
bash "$MDC" verify ornith:9b >/dev/null 2>&1; echo $?)
[[ "$rc" -eq 2 ]] && pass "dev: swapped override detected as MISMATCH (rc=2)" \
|| fail "dev: swapped override should MISMATCH (rc=$rc)"
# 3) SEC-14 core (prod): a FORGED override equal to the pin must NOT be accepted.
# Override is ignored; live backend is unreachable → cannot verify (rc=3), never OK.
err="$WORK/prod.err"
rc=$(set +e; CASAN_PROFILE=prod OLLAMA_HOST="$DEAD_OLLAMA" \
CASAN_MODEL_DIGEST_PIN="$PIN" CASAN_MODEL_DIGEST="digest-approved" \
bash "$MDC" verify ornith:9b >/dev/null 2>"$err"; echo $?)
[[ "$rc" -ne 0 ]] && pass "prod: forged override not accepted as OK (rc=$rc)" \
|| fail "prod: forged override was accepted (rc=0) — env override still trusted"
grep -q "MODEL_DIGEST_OVERRIDE_IGNORED" "$err" \
&& pass "prod: override explicitly ignored (MODEL_DIGEST_OVERRIDE_IGNORED)" \
|| fail "prod: override was not ignored (no OVERRIDE_IGNORED marker)"
# 4) strict flag: same enforced behavior without a full prod profile.
rc=$(set +e; CASAN_MODEL_DIGEST_STRICT=1 OLLAMA_HOST="$DEAD_OLLAMA" \
CASAN_MODEL_DIGEST_PIN="$PIN" CASAN_MODEL_DIGEST="digest-approved" \
bash "$MDC" verify ornith:9b >/dev/null 2>&1; echo $?)
[[ "$rc" -ne 0 ]] && pass "strict: forged override not accepted (rc=$rc)" \
|| fail "strict: forged override accepted (rc=0)"
echo ""
echo "===== SEC-14 SUMMARY: PASS=$PASS FAIL=$FAIL ====="
[[ "$FAIL" -eq 0 ]] || exit 1