feat: Skills extension phase 3 — CLI, TUI, directory reads, frontmatter check

Closes #2248.

Completes SEP-2640 support across all three clients, and closes the one
obligation #2234 deliberately left open.

- `checkSkillFrontmatterMatch` compares a served SKILL.md's own YAML
  frontmatter against the entry the listing advertised, field by field. No
  digest can cover this: a digest is taken over the bytes the server served,
  so it proves the file was not altered in transit and says nothing about
  whether the listing described it honestly. Needs a real YAML parser; `yaml`
  was already a root dependency, so this adds no package — but it does newly
  put it on core/'s import graph, hence the three bundler `external` lists.

- `resources/directory/read`: result schemas defined against the normative
  text, `InspectorClient.readResourceDirectory`, a Directory section on the
  Skills screen, and `--method resources/directory/read` in the CLI. The
  client refuses the call locally when the server did not declare
  `directoryRead`, which is the SEP's MUST NOT. A child the directory lists
  but the entry does not is marked `not listed` rather than merged into the
  manifest — the SEP calls a directory result a live observation and forbids
  treating it as extending the manifest.

- CLI: `skills/list`, `skills/get` and `resources/directory/read`, plus
  `--verify` — one NDJSON report per skill, a summary on stderr, exit 7 on a
  violation. Reads are sequential; a conforming manifest may declare 512
  entries.

- TUI: a Skills pane, shown only when the server declares the extension. Each
  row carries its conformance verdict as a glyph as well as a colour, since
  the pane is read over ssh and through script(1). Enter verifies.

- `verifySkills` re-throws `AuthRecoveryRequiredError` rather than recording
  it per file: it says the session's authorization expired, so absorbing it
  would report N identical read failures and swallow the error the TUI and
  the web commands key off to reauthorize.

Two open questions settled, both in code comments where they will be found:
`skills/get` carries no caching attributes because SEP-2640 leaves the
question open in as many words, so requiring them would fail a conforming
server; and there is no `PagedSkillsState` because every consumer of this
list is a whole-catalog verdict — computed over page one of three, "this
server's skills conform" is not merely partial but wrong.

The fixture grows two skills, both for checks that were otherwise
undemonstrable: `lying-listing` (listing and file disagree, digest still
verifies) and `stale-manifest` (serves a file its manifest does not declare).

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01ND46Ci64DC3tzBy9aG81Nc
Signed-off-by: cliffhall <cliff@futurescale.com>
This commit is contained in:
cliffhall 2026-09-07 18:10:37 -04:00
commit 9a8488681e

View file

@ -76,13 +76,14 @@ The TUI provides terminal-native tabs and panes for interacting with your MCP se
- **Resources**: Browse and read resources exposed by the server.
- **Prompts**: List and test prompts.
- **Tools**: View available tools and execute them with form-like inputs. A tool whose advertised schema carries a portability problem is flagged in the list — red `!` for a construct a shipping MCP client refuses, yellow `?` for one handled unevenly — and the detail pane lists each finding under **Schema Portability** with the path, the problem, and a concrete fix. The verdict comes from [`core/json/schemaLint.ts`](../../core/json/schemaLint.ts), shared with the web Tools tab and the CLI's `--strict` report, so the three cannot disagree ([#1005](https://github.com/modelcontextprotocol/inspector/issues/1005)).
- **Skills**: Shown only when the connected server declares the SEP-2640 Skills extension (`io.modelcontextprotocol/skills`), since it is a *server* declaration and so only knowable after connecting. The list marks each skill with its structural verdict — `✓` conforms, `!` warnings only, `✗` an error — using a glyph as well as a colour, because this pane is read over ssh, in tmux and through `script(1)`. The detail pane shows the entry's URI, description, conformance findings and manifest. **Enter** verifies the selected skill: one `resources/read` per manifest file, each hashed against its advertised digest, plus the frontmatter cross-check that compares the served `SKILL.md`'s own frontmatter against the one the listing advertised. Verification is a gesture rather than a page load because SEP-2640 says hosts MUST NOT retrieve a skill's files ahead of need. The checks are the same ones the web Skills tab and the CLI's `--verify` run ([#2234](https://github.com/modelcontextprotocol/inspector/issues/2234), [#2248](https://github.com/modelcontextprotocol/inspector/issues/2248)).
- **Protocol**: View JSON-RPC request/response/notification history (matches the web Protocol monitor).
- **Network**: View HTTP fetch traffic for SSE / Streamable HTTP servers (matches the web Network monitor).
- **Console**: View stdio stderr from the connected server process (matches the web Console monitor).
## Navigation
- Use the **Arrow Keys** (Left/Right) or **Tab** to switch between the main tabs (Resources, Tools, Prompts, etc.).
- Use the **Arrow Keys** (Left/Right) or **Tab** to switch between the main tabs (Resources, Tools, Prompts, Skills, etc.).
- Use the **Arrow Keys** (Up/Down) to scroll through lists of items.
- Press **Enter** to select an item, execute a tool, or fetch a resource.
- Press **Escape** or `Ctrl+C` to exit the application.