Add the mechanism section to the Windows rung Agent Note: identity routes (restricted-user, AppContainer) default-deny everything from a fresh SID and pay per-path ACE writes for every read, while the WRITE_RESTRICTED token keeps the caller's SID and double-checks write-class access only — reads pass on the normal check alone. Covers DISABLE_MAX_PRIVILEGE/LUA_TOKEN as the token-side limited-user synthesis and why SidsToDisable read restriction is unused.
Add the implemented Agent Note owning the win32 rung decision: mxc rejected (OS floor 24H2, BaseContainer tier only on 25H2+ with the OS feature enabled, and arbitrary-path reads demand wholesale host DACL writes), AppContainer rejected (no ambient read access — arbitrary-path reads unsupported), landstrip already rejected and AppContainer-shaped. Link the note from the backend README pair and restate the choice in the PR description.
JSDoc every sandbox-windows-acl export (the 138 verify-export-jsdoc violations); fix the sandbox-windows-acl README code block (undeclared workspaceRoot), bring its stale claims current (the provider wiring and kill-on-close are live), add the canonical Model Experience sections to both new packages, audited the backend's indirect Model Experience sentence, and regenerate docs/config-catalog.md.
The placer fixtures derive the basename with split('/'), which never splits a Windows absolute path (backslashes), so the two image-placement tests fail on win32 hosts. The production rewrite passes absolute paths through; the fixtures now accept both separators.
The Windows platform layer re-enables the base fs-sandbox row (removing its disable), but still inserted dsh-fs-local: both extend FileSystem and provide ctx.fs, so every shipped win32 profile failed at load. Delete the insert; fs-sandbox stays the single fs provider exactly as on POSIX. Sync the roster specs, the base/reference README pairs, the sandbox core doc (read-only grants no sink on Windows), the windows-shell JSDoc, and re-record the i18n pairings.
Rebuild of the fs-overwrite-diff-bound branch on current master. Adds the
diffBasisMaxBytes Config field (10 MiB default, capped by runtime
allocation/decode limits), gates both overwrite sides, and reads the prior
basis from the bounded opened descriptor in cancellation-aware chunks; any
post-stat size change returns a null basis. Also pins the one-extra-byte
growth probe with a regression and drops the now-covered v8 ignore.
Bilingual README updates for the four touched packages (ui-skill's claim
flow and deterministic-injection model experience, the apiproxy skills
domain, the shared renderSkillContent seam export, the catalog stitch
sentence), the implemented Agent Note triplet recording the decision and
its peer-product evidence, and the regenerated catalogs/graphs.
turnAgentFor owns the addressed-agent resolution and the model-unavailable
refusal both turn-starting methods repeat; the duplication gate flagged the
copied block.
The policy scenario now expects the user-only quadrant in the menu with
its marker (riding the description — the hint field is claim-state ghost
text, which the menu never renders), and a new skill-user-invoke scenario
drives /name args through the composer against the real host: the claim
lands skill.invoke, the transcript shows the dedicated card with the
collapsed <skill_content> body, and a paced replay answers the injected
turn deterministically.
Confine Windows command execution through a WRITE_RESTRICTED token whose restricting SIDs carry an orphan-SID write allowlist, ported from https://github.com/huoyaoyuan/windows-acl-restrict-poc (@ 10e4dfb). Every Win32 call is checked and fails closed - the POC silently ran children with the FULL token when CreateRestrictedToken failed.
- @deepseek-ai/dsh-sandbox-windows-acl: koffi primitives verified against the MinGW Windows headers (verify/abi-probe.cpp) plus the confinement runner ([node, runner, --workspace, --temp, --mode, --, argv...]: kill-on-close job, stdio passthrough, exit-code mirroring, windows-acl-run: failure signature, grant revocation). read-only = strict zero grants (NUL device not writable; documented). Windows-only execution: exempted from the Linux coverage lane (windowsOnlyCoverageExclusions).
- @deepseek-ai/dsh-sandbox-local: PLATFORM_CHAINS.win32 filled with the windows-acl runner (full enforcement, ACL denial dialect, runner-failure rules).
- @deepseek-ai/dsh-pwsh-sandbox: sandbox-consuming pwsh executor (call-for-call mirror of dsh-bash-sandbox) over a new argv-level seam in dsh-pwsh-local; per-file coverage complete via the fake-provider spec.
- bundle/base: the Windows platform layer mounts the confined pwsh roster - sandbox/policy/fs-sandbox/permission/approval re-enabled, the POSIX bash stack stays disabled.
Co-authored-by: Huo Yaoyuan <huoyaoyuan@hotmail.com>
The user guide's model-catalog section teaches the three new knobs at
task altitude — declare levels per model, pick the reasoning dialect,
reshape catalog models with modelOverrides — with the settings.yaml
example exercising all of them, plus an UNSUPPORTED_REASONING_EFFORT
troubleshooting row. The generated plugin config catalog picks up the
new Config fields, and the bilingual Agent Note records the decision,
the alternatives considered, and the schemastery materialization
constraint that chose false over {} as the disable spelling.
A user/message carrying the skill-invocation source materializes as its
own conversation node (name/args lifted off the source metadata, never
re-parsed from the body) and renders as a right-aligned bubble: the /name
chip plus the user's trailing text, with the injected <skill_content>
collapsed behind a disclosure. A record with an unreadable name degrades
to the injected-context row.
A menu pick or an entered /name line now claims the composer into an
args-tolerant skill.invoke transaction instead of shipping the literal
text and hoping the model loads the skill. This gives every user-invocable
skill a deterministic entry point — including disable-model-invocation
skills the catalog never shows the model (issue #1470). Candidates carry
a user-only hint, and the unreached legacy <skill> reference codec is
removed (decision 21 removal cut).
A route's modelOverrides dict customizes individual installed-catalog
models — key = catalog model id, value = the same fields a models entry
takes — while the rest of the catalog keeps serving, which a models list
cannot express because declaring one replaces the served set. An
override becomes the catalog entry's configuration and resolves through
the existing entry path, so capacities, reasoningEfforts, compat, and
request-default semantics are identical to a models entry's. Unlike
Pi's config layer, which ignores unknown ids, every override that lands
nowhere is refused at the write that produced it: beside a models list,
on a hand-declared route, naming a model the catalog does not describe,
or smuggling an id through the schema's unknown-key tolerance.
Both catalog renderings now tell the model that a directly invoked skill
arrives as an inline <skill_content> block to follow without re-loading it
through the skill tool — the seam rule that keeps the user-explicit path
and the model-autonomous path from double-injecting one skill.
A model entry's reasoningEfforts dict declares its selectable thinking
levels — key = offered level, value = the wire spelling dispatch sends;
only off may leave the value empty (supported, send nothing). false
strips reasoning from a catalog model; every level is materialized
explicitly into pi-ai's thinkingLevelMap so nobody has to know pi-ai's
asymmetric absent-key defaulting. compat.thinkingFormat and
compat.supportsReasoningEffort become configurable on the route and per
model (model > route > catalog entry > pi-ai's URL-derived guess),
openai-completions only, so a private gateway speaking the DeepSeek
reasoning dialect no longer depends on its URL being recognizable.
Record-typed drift gates pin both enums to pi-ai's, and an unserviceable
declaration is refused at the write that produced it, naming route,
model, and level.
skill.list now serves every user-invocable skill and carries modelInvocable
so menus can mark user-only entries; the old model-and-user intersection
hid disable-model-invocation skills from their only legitimate entry point
(issue #1470). skill.invoke enforces user-invocation policy at the host
boundary, renders the canonical <skill_content> body, and injects it as a
user-role message carrying the skill-invocation source before starting a
turn. The connection fixture mirrors both faces for client tests.
The model-facing <skill_content> rendering moves from dsh-tool-skill to the
dsh-skill seam so the skill tool result and the upcoming user-explicit
invocation injection share one canonical shape. The seam also declares the
skill-invocation MessageSource kind that injection will stamp on its
user-role messages.
imccyu
pku-xht
ZiyaZhang
Huanqi Cao
Yichen Jiang
Tianyi Cui
