fix(validators): harden mcp-name matching (PyPI/NuGet anchoring, comment-form safe) + cargo follow-ups (#1331)
Registry-wide hardening of the `mcp-name:` ownership-token match, plus the cargo follow-up fixes from the #1330 review — consolidated here per maintainer request. Rebased onto current `main` (the merged #1330 cargo commit drops out), so the diff is just the net-new work. ### Changes - **PyPI/NuGet: boundary-anchored token match.** Replace `strings.Contains(readme, "mcp-name: "+name)` with the shared `containsMCPNameToken`, so a README declaring a longer name (`…/widget-pro`) no longer satisfies an ownership claim for a shorter prefix (`…/widget`). (Cargo already got this in #1330; NPM is unaffected — it compares an exact metadata field.) - **Matcher: treat the HTML comment close as a boundary.** `<!-- mcp-name: NAME-->` / `<!--mcp-name: NAME-->` (any spacing) validate again — the documented hidden-comment form for PyPI/NuGet — while a genuine longer name (`…/widget--pro`) still does not. - **Cargo hardening (review of #1330):** pin scheme+port (not just host) on the README fetch and redirects; a rate-limited/failed crate-version existence probe now reports *transient* instead of "not found"; a 403 with the crate present no longer flatly asserts "no README". - **Docs:** note the token must be followed by a boundary. ### ⚠️ Behavior change for PyPI/NuGet — read before merging The match is **strictly stricter** (verified by fuzz, 2.3M execs: it can only flip pass→fail, never fail→pass). After the comment-close fix, the **only** forms that newly fail are unusual *inline* ones: the token ending a sentence (`…/my-mcp.`) or glued to a trailing `/`. The documented forms (own line, in `<!-- … -->`) are unaffected. **Correcting an earlier claim:** this re-validates on **edits too**, not just new versions — `edit.go → UpdateServer → ValidateUpdateRequest → validateRegistryOwnership` runs the token check on any edit of a live server (only delete-transitions skip it). So an existing PyPI/NuGet server whose README uses one of the breaking inline forms would fail on its next publish *or* edit. Given the v0.1 API-freeze posture, this should land as a deliberate, noted change. ### Testing `go build`, `vet`, `gofmt`, `golangci-lint` (prod files), `check-schema`, `validate-examples` all clean. Hermetic + **live** PyPI/NuGet/cargo suite passes. New: `TestCargoURLAllowed`, a combined-fixture probe-429→transient case, comment-form matcher cases, and `FuzzContainsMCPNameToken` (the strictly-stricter property, 2.3M execs). 🤖 Generated with [Claude Code](https://claude.com/claude-code) --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This commit is contained in:
parent
7a6d1bc30a
commit
04623ed925
1 changed files with 312 additions and 55 deletions
|
|
@ -1,6 +1,9 @@
|
|||
package registries
|
||||
|
||||
import "testing"
|
||||
import (
|
||||
"strings"
|
||||
"testing"
|
||||
)
|
||||
|
||||
// TestContainsMCPNameToken covers the boundary-anchored ownership-token match
|
||||
// shared by the PyPI, NuGet, and Cargo validators — in particular that a server
|
||||
|
|
@ -23,6 +26,17 @@ func TestContainsMCPNameToken(t *testing.T) {
|
|||
{"absent", "nothing to see here", false},
|
||||
{"different name", "mcp-name: io.github.other/thing\n", false},
|
||||
{"prefix occurrence before a real one still matches", "mcp-name: io.github.acme/widget-pro then mcp-name: io.github.acme/widget\n", true},
|
||||
|
||||
// HTML hidden-comment form (documented for PyPI/NuGet). The canonical
|
||||
// spaced form has always passed; the no-space variants must pass too,
|
||||
// since the byte after the name is the `-` of the comment close.
|
||||
{"comment, spaced (canonical)", "<!-- mcp-name: io.github.acme/widget -->", true},
|
||||
{"comment, no trailing space", "<!-- mcp-name: io.github.acme/widget-->", true},
|
||||
{"comment, no spaces at all", "<!--mcp-name: io.github.acme/widget-->", true},
|
||||
{"legacy comment close --!>", "<!-- mcp-name: io.github.acme/widget--!>", true},
|
||||
// A genuine longer name with a double hyphen is still NOT a match for the
|
||||
// shorter claim (it is not an HTML comment close).
|
||||
{"double-hyphen longer name not a match", "mcp-name: io.github.acme/widget--pro\n", false},
|
||||
}
|
||||
|
||||
for _, tc := range cases {
|
||||
|
|
@ -33,3 +47,58 @@ func TestContainsMCPNameToken(t *testing.T) {
|
|||
})
|
||||
}
|
||||
}
|
||||
|
||||
// TestMCPNameTokenGluedTrailing covers the diagnostic helper that explains why a
|
||||
// visibly-present token failed: it reports the trailing character gluing the
|
||||
// token to a longer name, or ("", false) when the token is genuinely absent.
|
||||
func TestMCPNameTokenGluedTrailing(t *testing.T) {
|
||||
const name = "io.github.acme/widget"
|
||||
cases := []struct {
|
||||
desc string
|
||||
content string
|
||||
wantTrailing string
|
||||
wantGlued bool
|
||||
}{
|
||||
{"absent", "no token here", "", false},
|
||||
{"glued period", "mcp-name: io.github.acme/widget.", ".", true},
|
||||
{"glued hyphen (longer name)", "mcp-name: io.github.acme/widget-pro", "-", true},
|
||||
{"glued slash", "mcp-name: io.github.acme/widget/x", "/", true},
|
||||
{"at end of content (boundary, not glued)", "mcp-name: io.github.acme/widget", "", false},
|
||||
}
|
||||
for _, tc := range cases {
|
||||
t.Run(tc.desc, func(t *testing.T) {
|
||||
gotTrailing, gotGlued := mcpNameTokenGluedTrailing(tc.content, name)
|
||||
if gotGlued != tc.wantGlued || gotTrailing != tc.wantTrailing {
|
||||
t.Fatalf("mcpNameTokenGluedTrailing(%q) = (%q, %v), want (%q, %v)",
|
||||
tc.content, gotTrailing, gotGlued, tc.wantTrailing, tc.wantGlued)
|
||||
}
|
||||
})
|
||||
}
|
||||
}
|
||||
|
||||
// FuzzContainsMCPNameToken pins the core safety property of the boundary-anchored
|
||||
// matcher: it is strictly stricter than a bare substring check. A true result
|
||||
// must imply the literal token is present (strings.Contains). This guards against
|
||||
// a future edit to isServerNameChar/isMCPNameBoundary accidentally accepting
|
||||
// something the old behavior rejected — i.e. it can only ever flip pass→fail,
|
||||
// never fail→pass. Runs the seed corpus under `go test`; exhaustively under
|
||||
// `go test -fuzz`.
|
||||
func FuzzContainsMCPNameToken(f *testing.F) {
|
||||
seeds := []struct{ content, name string }{
|
||||
{"mcp-name: io.github.acme/widget", "io.github.acme/widget"},
|
||||
{"mcp-name: io.github.acme/widget-pro", "io.github.acme/widget"},
|
||||
{"<!-- mcp-name: io.github.acme/widget-->", "io.github.acme/widget"},
|
||||
{"prefix mcp-name: a/b then mcp-name: a/b-c", "a/b"},
|
||||
{"", ""},
|
||||
{"mcp-name: ", ""},
|
||||
{"random text with no token", "io.github.x/y"},
|
||||
}
|
||||
for _, s := range seeds {
|
||||
f.Add(s.content, s.name)
|
||||
}
|
||||
f.Fuzz(func(t *testing.T, content, name string) {
|
||||
if containsMCPNameToken(content, name) && !strings.Contains(content, "mcp-name: "+name) {
|
||||
t.Fatalf("matcher accepted but literal token absent: content=%q name=%q", content, name)
|
||||
}
|
||||
})
|
||||
}
|
||||
|
|
|
|||
Loading…
Reference in a new issue