Compare commits

..
7 changed files with 36 additions and 96 deletions
+9 -26
View File
@@ -39,18 +39,6 @@ keeps every weekly edition looking the same regardless of which model writes the
- **Security analysis.** Scan the per-recipe `upgrade_notes_md` + the summary (and use your own
knowledge of the version bumps) for upgrades that fix **CVEs / security issues**. For each recipe,
**count the CVEs** the PR fixes — this drives both the table's `cve` column and the priority sort.
Before crediting a CVE to an open PR, compare that PR's head `compose.yml` with the **current
target-branch tip**, both pinned by SHA at report time. Run
`advisory-scan.py <recipe> --compose-from <current-target-sha-raw-url>
--compose-to <head-sha-raw-url>` when an upgrade log lacks that exact scan, or when its listed
image windows disagree with the PR diff. Credit image CVEs only for image version changes in
that diff, and only if the PR head is the fixed version. An upgrade-time merge base may now be
stale if upstream advanced; a previous release or warm canonical can also include fixes already
present in the target branch. For example, lasuite-docs PR #11 did not change Valkey,
so its 8.1.11→9.1.3 CVE must not be credited to that PR. Record a security-relevant config
fix separately with direct evidence that the target branch lacks it. If the current diff cannot
be verified, use `?` and explain why. If the PR head is behind the target branch or unmergeable,
say so in the same row; its old CI result is not proof that the current combined recipe works.
- **ADDITIONALLY, and never instead:** each per-recipe log carries an `### Advisory scan
(deterministic pre-step)` block (from `cc-ci-plan/advisory-scan.py` — GitHub Security
Advisories + vendor security pages + OSV, with severities and fixed-in versions). Treat its
@@ -60,13 +48,10 @@ keeps every weekly edition looking the same regardless of which model writes the
lists **failed sources**, the count is **not** authoritative: render the cve cell as `?`
(unknown), never `none` — a blank that reads as "clean" is exactly how two CVSS-9.8 gitea
RCEs were reported as "none" on 2026-08-07.
- The table's **CVEs** column counts known CVEs fixed by the **current open PR's diff against
current target branch**, even if `/upgrade-all` found no new image version this week. Keep the
existing one-row-per-recipe table and show that PR in the row. `SKIPPED — up-to-date` describes
this week's image survey, not the absence of a relevant open PR; use the PR's actual CI verdict
in TESTS when one exists. A zero means the current PR fixes no verified CVEs (or there is no
PR); it does not certify that deployed images are vulnerability-free. If the PR's advisory
scan is absent, a source failed, or an image window was unmatched, use `?`.
- The table's **CVEs** column counts known CVEs addressed by the available recipe upgrade.
A zero for an up-to-date recipe means there is no upgrade window to credit; it does not
certify that the running images are free of vulnerabilities. If an upgrade exists but its
advisory scan is absent, a source failed, or an image window was unmatched, use `?`.
If a confirmed security advisory is fixed but has no CVE ID, put `"YES"` in `cve` rather
than `0`; describe its advisory ID and severity in Notes. If CVE IDs are also fixed, show
the numeric CVE count and mention the additional non-CVE advisories in Notes.
@@ -85,10 +70,8 @@ keeps every weekly edition looking the same regardless of which model writes the
When a sidecar contributes a critical/high, name the image in the bulletin — CVE-2025-49844 is
a redis flaw, not a discourse one, and an operator reading "discourse" needs to know that.
(The scan headline itself now says `UNKNOWN` rather than a number in that case.)
In particular: **no new image upgrade this run does not force `0`** if an open PR still has
an unmerged security change. Recheck its current diff and credit only fixes absent from the
current target branch. A recipe with no relevant PR reports `0`; a clean PR scan reports its
number (including `0`). Benign notes in a scan
In particular: a recipe with **no upgrade this run** (up-to-date/skipped) has nothing an
upgrade could have fixed — report `0`, not `?`. A recipe with a clean scan reports its number (including `0`). Benign notes in a scan
block (`no-advisories-published`, `skipped: template URL`) are NOT failures and must not
trigger `?`. If you find yourself rendering `?` for many recipes, that is a bug to report in
the Addendum, not a normal outcome.
@@ -108,9 +91,9 @@ keeps every weekly edition looking the same regardless of which model writes the
re-runs), drift between the summary and reality, a **misleading branch name**, a tooling snag that
silently drops a recipe, leftover artifact PRs, or **something in the test run that could be
improved**. **Only real issues — if everything looks clean, leave it empty; do NOT invent items.**
- **What changed — per recipe with a relevant open PR.** Describe the PR's current diff (versions,
configuration, migrations/gotchas, reconcile notes), including an older PR still awaiting merge.
Say when it was last updated and whether it can merge cleanly. Skip recipes with no relevant PR.
- **What changed — per recipe with a PR.** One short section per recipe that has a PR this week
describing what the upgrade actually changes (versions, notable deps, migrations/gotchas, reconcile
notes). Skip recipes with no PR.
4. **Write the spec** `/tmp/report-spec.json` (shape in the helper header):
- `date`, `subtitle` ("Week of <human date>"),
+16 -11
View File
@@ -159,21 +159,26 @@ On cc-ci's `~/.abra/recipes/<recipe>` (wrap every abra call per the pseudo-TTY b
### 2a. Advisory scan (deterministic; ADDITIVE — run it, never skip it)
Run the deterministic scanner against the **exact proposed PR diff** and paste its markdown block
verbatim into the per-recipe log. Use the proposed head's `compose.yml` and the commit it branches
from (the merge base with the target branch), both pinned by commit SHA:
Run the deterministic scanner for the exact upgrade window and **paste its markdown block verbatim
into the per-recipe log**:
```
python3 /srv/cc-ci/cc-ci-plan/advisory-scan.py <recipe> \
--compose-from <raw-compose-url-at-pr-base-sha> \
--compose-to <raw-compose-url-at-pr-head-sha>
python3 /srv/cc-ci/cc-ci-plan/advisory-scan.py <recipe> --from <old-app-version> --to <new-app-version> \
[--image <name>=<old>:<new>]...
```
Do not substitute the previous published release, warm canonical, or a hand-typed image window for
the PR base. Those answer a different question and can credit a PR for CVEs fixed before its branch
was created. The compose diff includes every changed image, including sidecars. If the PR changes no
image tags, record zero image-version CVEs for this PR; assess any security-relevant configuration
change separately, with its own evidence.
**Pass an `--image` for EVERY sidecar you upgraded** (redis, postgres, nginx …), not just the app —
each image is judged by its own versions, and an image you don't name is not counted at all. Repeat
the flag for each one and pass them **all in a single call** (the count is a union across images).
e.g. discourse moving app 3.5.3→2026.7.1 *and* redis 7.4→8.10:
```
... --from 3.5.3 --to 2026.7.1 --image redis=7.4:8.10
```
→ 140 CVEs (123 app + 17 redis), where the redis seventeen include a **critical** (CVE-2025-49844)
that is invisible if the sidecar is left out. `<name>` is substring-matched against source repo names,
so make it specific enough to hit exactly one.
**If the scan reports advisories it could NOT judge, re-run it with `--adjudicate`.** That is a second
pass: it collects each open case's full evidence (advisory prose, references, affected ranges, every
-5
View File
@@ -56,11 +56,6 @@ advisory-scan.py <recipe> --compose-to <URL> [--compose-from <URL>] # windows
Exit code is always 0 — this is informational. Failures are *reported*, never raised.
For PR attribution, pin both compose URLs to commits: the PR head and its merge base with the
target branch. A previous release or warm canonical can include image changes already present
before the PR. A compose diff with only sidecar changes has no primary app window; every changed
sidecar is scanned under its own image name.
---
# Pass 1 — deterministic
+6 -11
View File
@@ -1336,8 +1336,7 @@ def _default_branch_compose(url: str) -> str | None:
return f"{host}/{owner}/{repo}/raw/branch/{br}/{path}" if br else None
def windows_from_compose(to_url: str, from_url: str | None = None,
include_primary: bool = False) -> tuple:
def windows_from_compose(to_url: str, from_url: str | None = None) -> tuple[list, str | None]:
"""Derive the scan's version windows by DIFFING two compose files.
This is the deterministic alternative to a human (or a model) deciding which `--image` args a
@@ -1367,8 +1366,7 @@ def windows_from_compose(to_url: str, from_url: str | None = None,
else:
others.append((repo.split("/")[-1], prev_tag, tag))
wins = ([app] if app else []) + others
result = (wins, f"baseline {from_url}")
return (*result, bool(app)) if include_primary else result
return wins, f"baseline {from_url}"
def main() -> int:
@@ -1397,8 +1395,7 @@ def main() -> int:
a = ap.parse_args()
images = []
if a.compose_to:
wins, note, app_changed = windows_from_compose(a.compose_to, a.compose_from,
include_primary=True)
wins, note = windows_from_compose(a.compose_to, a.compose_from)
if not wins:
print(f"### Advisory scan — {a.recipe}\n\n**No image versions changed between the two "
f"compose files, so this upgrade fixes no CVEs by definition.**\n\n_{note}_")
@@ -1406,13 +1403,11 @@ def main() -> int:
print(f"_derived from compose diff ({note}):_", file=sys.stderr)
for n_, f_, t_ in wins:
print(f"_ {n_}: {f_} → {t_}_", file=sys.stderr)
# Only an actual app-service change defines the primary window. A sidecar-only PR must
# not be scanned as though its first sidecar were the application image.
if app_changed:
# The `app` service (first entry when present) drives --from/--to; the rest are --image
# windows. Passing every window as --image too is harmless: each is matched by name against
# the advisory sources, and an unmatched name is simply ignored.
a.v_from, a.v_to = a.v_from or wins[0][1], a.v_to or wins[0][2]
images = list(wins[1:])
else:
images = list(wins)
for spec in a.image:
name, _, rng = spec.partition('=')
vf, _, vt = rng.partition(':')
+1 -5
View File
@@ -100,11 +100,7 @@ def survey(date):
if s.get("context") == "cc-ci/testme":
ci = {"state": s.get("state"), "url": s.get("target_url"), "desc": s.get("description")}
rec["open_prs"].append({"number": p["number"], "title": p["title"],
"url": p["html_url"], "head": sha,
"base": p.get("base", {}).get("sha"),
"created_at": p.get("created_at"),
"updated_at": p.get("updated_at"),
"mergeable": p.get("mergeable"), "ci": ci})
"url": p["html_url"], "head": sha[:10], "ci": ci})
# Per-recipe upgrade plan/summary (has the breaking-change + CVE/security analysis) — the agent
# mines this to lead with critical-CVE upgrades.
nf = os.path.join(LOGDIR, "upgrades", f"{r}-upgrade-{date}.md")
-35
View File
@@ -746,41 +746,6 @@ volumes:
def test_no_change_yields_no_windows(self):
self.assertEqual(self._windows(old=self.NEW, new=self.NEW), [])
def test_lasuite_docs_pr_does_not_claim_preexisting_valkey_upgrade(self):
base = """services:
app:
image: lasuite/impress-frontend:v6.0.4
yhub-valkey:
image: valkey/valkey:9.1.3-alpine
docspec:
image: ghcr.io/docspecio/api:3.0.2
"""
head = base.replace("ghcr.io/docspecio/api:3.0.2", "ghcr.io/docspec/api:1.21.4")
windows = self._windows(old=base, new=head)
self.assertEqual(windows, [("api", "3.0.2", "1.21.4")])
self.assertNotIn("valkey", [name for name, _, _ in windows])
captured = {}
def fake_scan(recipe, v_from, v_to, registry, images):
captured.update(from_version=v_from, to_version=v_to, images=images)
return {"recipe": recipe, "from": v_from, "to": v_to,
"count_known": True, "cve_count_fixed": 0, "cves": {}, "sources": [],
"fixed_by_this_upgrade": [], "unclassified": [], "registry_urls": 0,
"sources_failed": [], "sources_benign": [], "windows": {},
"classified_by": {}}
with unittest.mock.patch.object(A, "windows_from_compose",
return_value=(windows, "pinned base", False)), \
unittest.mock.patch.object(A, "scan", fake_scan), \
unittest.mock.patch.object(sys, "argv", ["advisory-scan.py", "lasuite-docs",
"--compose-from", "base", "--compose-to", "head"]), \
unittest.mock.patch.object(sys, "stdout", io.StringIO()), \
unittest.mock.patch.object(sys, "stderr", io.StringIO()):
A.main()
self.assertIsNone(captured["from_version"])
self.assertEqual(captured["images"], windows)
def test_templated_tags_are_skipped(self):
new = self.NEW.replace('ghcr.io/plausible/community-edition:v3.2.1', 'ghost:${IMAGE_VERSION}')
self.assertNotIn("ghost", [n for n, _, _ in self._windows(new=new)])
+2 -1
View File
@@ -4,12 +4,13 @@
|---------|-------|-------------|----------------------|
| app | nginx | https://github.com/nginx/nginx | https://nginx.org/en/CHANGES |
| git | alpine/git | https://github.com/alpine-docker/git | https://hub.docker.com/r/alpine/git/tags |
| ssh | lscr.io/linuxserver/openssh-server | https://github.com/linuxserver/docker-openssh-server | https://github.com/linuxserver/docker-openssh-server/releases |
## Standing notes
- nginx even-numbered minor versions (1.28.x, 1.30.x) are mainline; odd (1.27.x, 1.29.x) are stable. The recipe tracks mainline.
- nginx 1.31.7 (14 Oct 2026) is a security patch release with FIVE security fixes: CVE-2026-15176 (buffer overflow in a worker using `alias` when the URI was changed by `rewrite`), CVE-2026-93479 (buffer overflow evaluating string expressions with variables — affects document-root calculation and request logging in http/stream), CVE-2026-95101 (HTTP/3: part of an unknown frame treated as request body — bypasses `client_max_body_size` / injects into proxied request), CVE-2026-97629 (TLS session-ticket key rotation left one key uninitialized — session forgery), CVE-2026-97205 (use-after-free with async file I/O when an SSI page has `include virtual`). Features: geoip MMDB support, backend SSL cert verification by IP, QUIC reconfiguration + binary upgrade with `quic_bpf`. Change: `ngx_http_v2_module` now uses RFC 9218 prioritisation (RFC 7540 priority signals ignored). Bugfixes (X-Accel-Redirect cached response segfault, dav COPY/MOVE with alias segfault, QUIC flood false positives, SSL close_notify handling). NONE affect the custom-html static-file-serving use case (no HTTP/3/QUIC, no AIO/SSI, no `rewrite`+`alias` path mapping, no session-ticket rotation config) — but the 5 CVEs are why this bump is warranted. No breaking config changes; no operator action beyond the image pull.
- compose.git-pull.yml is an optional overlay for git-pull functionality; alpine/git version lives there.
- compose.sftp.yml and compose.sso.yml are other optional overlays; linuxserver/openssh-server uses `latest` tag and is not version-pinned.
- compose.sftp.yml and compose.sso.yml are other optional overlays; linuxserver/openssh-server uses the moving `latest` tag and is pinned by digest (`image:latest@sha256:…`). Because `latest` is rebuilt regularly, its upstream multi-arch index digest moves even when no tag/version changes; re-check the index digest directly with `docker buildx imagetools inspect lscr.io/linuxserver/openssh-server:latest --format '{{.Manifest.Digest}}'` (abra FATAs on `tag@digest` refs and cannot see this). 2026-11-13: pinned `aa4e1b39…` was stale; live index digest is `sha256:161e0b670b31d9539ca1718bc5943e981779331b4e1c9d32e8929d7652a1f22f` (amd64 child `7bd52bcb…`, arm64 `699500b1…`).
- nginx 1.31.6 (15 Sep 2026) is a security patch release: CVE-2026-90439 (heap buffer overflow in a worker under certain configurations when using HTTP/3 with OpenSSL 3.5.0 and earlier). Change: QUIC transport parameters extension received in an SSL connection is now always ignored. Bugfixes: binary upgrade with control API socket + perl module; predicate-location evaluation error ignored; nested location lookup with regex/predicate locations; segfault reading config with `geo` … `ranges` + corrupted binary base file. None affect the custom-html static-file-serving use case (no HTTP/3/QUIC, no predicate locations, no perl, no `geo`).
- Breaking change in nginx 1.31.0: HTTP/2 and HTTP/3 requests with Connection/Proxy-Connection/Keep-Alive/Transfer-Encoding/Upgrade/TE headers are now rejected.
- Breaking change in nginx 1.29.7: keepalive in upstream block is enabled by default; proxy_http_version changed to 1.1; Connection proxy header no longer sent by default.