* refactor: remove guard mode execution paths and guard-only packages
Guard (non-proxy) mode is removed; all package-manager commands now
always run the proxy flow. Removes the guard engine, the common flow,
the extractor package, the npm/pypi dependency resolvers and the
PackageResolver plumbing that only guard mode consumed.
The guard package retains only PackageManagerGuardInteraction, which
the proxy flow and confirmation interceptors reuse for user prompts.
Proxy behavior is unchanged.
* refactor: remove proxy opt-out surfaces, guard references in config, action and docs
Removes Config.ProxyMode, ProxyConfig.Enabled, IsProxyModeEnabled, the
proxy_mode legacy fallback, PMG_PROXY_ENABLED handling and the
--proxy-mode / --include-dev-dependencies flags. Proxy interception can
no longer be disabled.
Also removes the proxy-mode input from the GitHub Action, the
proxy-mode doctor check and setup info row, updates the E2E workflow to
stop passing --proxy-mode=false, and sweeps guard-mode wording from
docs and the config template.
The legacy proxy_install_only flat key and PMG_PROXY_INSTALL_ONLY env
var remain supported. audit.FlowTypeGuard is kept so previously
recorded audit events still translate for cloud sync.
* feat: fail loudly when a removed proxy opt-out is still configured
A leftover proxy.enabled: false / proxy_mode: false config key or
PMG_PROXY_ENABLED=false / PMG_PROXY_MODE=false env var previously meant
guard mode; silently ignoring it would switch those users to proxy
interception without notice. PMG now exits with an actionable error
naming the exact source. Precedence mirrors the old resolution order:
env (ignored under lockdown) > proxy.enabled > legacy proxy_mode.
The pmg config subtree is exempt so the config file can still be fixed
with pmg config edit/set. The GitHub Action's proxy-mode input is kept
as a tombstone that fails the action when set to false and warns
otherwise.
* refactor: extract flows.RunProxy and address review findings
Collapses the identical parse-then-run body duplicated across the 12
package manager commands into flows.RunProxy. Documents the cache-hit /
offline analysis trade-off versus the removed guard manifest path,
fixes a stale non-proxy label in the E2E workflow and a stale guard
reference in the uvx parser comment.
* fix(config): mirror old proxy opt-out precedence exactly
PMG_PROXY_MODE only ever took effect through the legacy fallback, which
was gated on the presence of a proxy: key in the config file (even a
null one). Promoting it to the top env tier caused two inversions: a
stale PMG_PROXY_MODE=false hard-failed configs that resolved to proxy
mode, and PMG_PROXY_MODE=true silently overrode an explicit
proxy.enabled: false file opt-out. The check now resolves in the old
order: PMG_PROXY_ENABLED > proxy: section (presence gates the legacy
tier) > PMG_PROXY_MODE > flat proxy_mode.
parseOptOutBool also accepts numeric values (0 = false) to match
viper's WeaklyTypedInput/cast.ToBool coercion, so proxy.enabled: 0 and
proxy_mode: 0 are detected as opt-outs.
* refactor: move package manager interaction out of guard
* refactor: trim package manager interaction
* fix(config): normalize config keys viper-style in proxy opt-out check
Viper resolved config file keys case-insensitively and expanded dotted
keys, so spellings like Proxy:, Enabled:, a literal proxy.enabled key or
Proxy_Mode selected guard mode before the removal. The opt-out check now
lowercases keys recursively and nests dotted keys before matching, so
those existing opt-outs fail loudly instead of being silently ignored.
* refactor: remove inert transitive controls, dead parser state and guard audit variant
transitive / transitive_depth lost their only consumers with the
dependency resolvers; remove the config fields, flags, template and doc
entries, and the report/audit plumbing that misreported transitive
analysis as enabled.
Remove write-only parser state (PackageInstallTarget.Extras,
ParsedCommand.ManifestFiles, ShouldExtractFromManifest); IsManifestInstall
stays as it feeds sandbox gating via IsInstallationCommand.
Remove audit.FlowTypeGuard and its cloud mapping; guard events recorded
by pre-removal versions in an unsynced WAL translate to UNSPECIFIED.
* fix: address review findings on the opt-out wiring and cleanups
Move the removed-opt-out rejection from the CLI PersistentPreRun into
proxyFlow.Run: the check now fires exactly for package-manager runs, so
non-install commands (pmg setup remove, doctor, config, version) stay
usable to fix or remove an opted-out installation, and future commands
inherit or avoid the check by construction instead of by exemption list.
Also: make the e2e malicious-package assertion actually fail the job
when an install is not blocked, route pmg go through flows.RunProxy,
and drop the dead extras return from pypiParsePackageInfo (extras are
still stripped from package names).
* fix(config): make the removed opt-out check faithful to the old resolution
The gate that silenced the legacy proxy_mode surfaces matched the raw
proxy key case-sensitively in the old code, while values resolved
viper-style (case-insensitive, dotted keys); applying each semantic
where the old code did fixes both divergences: a case-variant Proxy:
section no longer hides a flat proxy_mode: false opt-out, and a dotted
proxy.enabled: false overridden by proxy_mode: true no longer errors.
Replace the generic key-tree normalization with two targeted lookups
(the check only ever resolves proxy.enabled and proxy_mode), which also
makes colliding spellings resolve deterministically. Coerce legacy-tier
values cast.ToBool-style so PMG_PROXY_MODE=off style opt-outs are
detected, log the config read error instead of swallowing it, and
shorten the error to a one-line statement with the specific remedy in
the help text.
Add lockdown coverage (env inert both directions) and a repeated-run
determinism test.
* fix(config): fall back to defaults for unrecognized proxy opt-out values
The old loader swallowed viper errors and ran on defaults, so values
like proxy.enabled: yes or PMG_PROXY_ENABLED=banana silently discarded
the whole config and defaulted to proxy. Treat them the same way now:
unrecognized values mean the default (proxy on) instead of a hard
error, and the doc comment no longer claims the old loader failed
loudly. Only values that actually meant guard mode fail.
Also check the removed opt-out before the CA trust check in pmg go,
restoring the old error precedence: a config problem must not steer
the user into an unnecessary OS trust store change.
* fix(e2e): PMG_PROXY_MODE assertion must match the legacy gate semantics
The runner's setup step writes the template config, which has a proxy:
section — and with one present the legacy PMG_PROXY_MODE was always
inert, so expecting a loud failure there asserts pre-fidelity-fix
behavior. Assert both sides instead: inert (command succeeds) with the
standard config, loud failure against an empty config dir where the
legacy fallback actually applied.
* refactor(config): collapse parseOptOutBool to ParseBool over the string form
YAML hands us typed values (bool, int), so route them through
fmt.Sprintf %v and strconv.ParseBool instead of a per-type switch.
Identical behavior for every recognized value; numbers other than 0/1
now read as no opinion instead of cast.ToBool's nonzero-true, which no
real config relies on.
* feat(sandbox): introduce presets - additive workload allowance bundles
Presets are named, additive-only bundles of sandbox allowances for a
specific workload (git hooks tooling, Astro/Vite/Next.js dev servers).
They solve the per-workload tuning friction from #384 without weakening
the default posture: no built-in profile references a preset, presets
cannot carry deny rules or profile booleans (strict YAML decoding), and
mandatory denies still win everywhere except the existing exact-match
suppression.
- Preset schema with metadata (author, labels) and schema_version gating
- Registry over ordered sources (embedded builtin, user dir); builtin
wins name collisions; source abstraction is the extension point for a
future hosted registry and SafeDep cloud sync
- Official presets: git, astro, vite, nextjs (with threat notes)
- Overlay/runtime integration: pmg sandbox allow preset=<name> and
--sandbox-allow preset=<name>, stored by reference, resolved at apply
time, missing presets warn (fail closed) instead of aborting
- Profile integration: presets: [...] list resolved after inherits
- CLI: pmg sandbox preset list (metadata filters, --json), show (prints
YAML with threat notes), lint
- Docs: user guide (docs/sandbox-presets.md) and design spec
Closes#384
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PETCfE4crLcmodosz12qRn
* fix(sandbox): address review findings on presets
- Presets never modify deny lists: a profile authored deny now survives a
preset allowing the same path (deny-beats-allow keeps it enforced).
Regression test added.
- Profile inspection commands (show, diff, lint) construct the profile
registry with the user-aware preset registry so they agree with runtime
resolution of custom profiles referencing user presets.
- Handle stderr write error when warning about unresolvable presets.
- Compute preset show underline from the uncolored header.
- Use path.Join for embed.FS reads (slash-separated on all platforms).
- Clarify in docs that lint-staged/astro are examples of preset workloads.
- Drop the design spec from the PR per review.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PETCfE4crLcmodosz12qRn
* fix(sandbox): harden preset precedence against authored denies
Addresses external security review findings on the preset mechanism:
- Bubblewrap: a mandatory write-denied path listed in allow_read lost its
protection when a later writable parent bind covered it (bwrap last
mount wins) - exactly the git preset shape (allow_read .git/config +
allow_write .git/**). The mandatory deny now re-binds the path
read-only after all writable mounts instead of being skipped.
Regression test asserts mount ordering. Landlock and Seatbelt were
unaffected (tests added for the same policy shape on Landlock).
- Environment: ScrubEnv is allow-wins, so a preset environment allowance
could override a profile-authored deny. Preset env allowances
overlapping an authored deny pattern are now dropped at application
time (conservative bidirectional glob overlap, fail closed). Surviving
entries still opt out of built-in credential scrubbing as intended.
- Network: removed allow_outbound from the preset schema. Both platform
translators are all-or-nothing for outbound (one allow rule means
blanket network access), so a preset outbound entry would silently
change network posture far beyond what its YAML conveys. Strict
decoding rejects the key.
- Added a dual-path expansion equivalence test (profile presets: field
vs overlay/--sandbox-allow) and documented the precedence guarantees
in docs/sandbox-presets.md.
Explicit --sandbox-allow and pmg sandbox allow overrides keep their
existing semantics.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PETCfE4crLcmodosz12qRn
* docs(sandbox): document env and preset allowances in allow command and overlay docs
pmg sandbox allow help, the --sandbox-allow flag usage, and the project
overlay docs enumerated only read/write/exec/net types. Add env and
preset to all of them, with an overlay example for persisting an env
allowance and a note on why env entries are not auto-promoted by --last.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PETCfE4crLcmodosz12qRn
* chore(sandbox): trim preset code comments to corner cases and minimal godocs
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PETCfE4crLcmodosz12qRn
* fix(sandbox): exact glob intersection for preset env deny overlap
The bidirectional literal-text heuristic missed overlapping globs with
different literal structure: preset allow AWS_*_KEY and authored deny
AWS_SECRET_* both match AWS_SECRET_ACCESS_KEY but neither pattern
matches the other's text, so the allowance merged and allow-wins
scrubbing exposed the variable. EnvPatternsOverlap now computes exact
intersection non-emptiness for the name glob dialect (case-insensitive,
'*' any sequence, '?' single char) via memoized DP.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PETCfE4crLcmodosz12qRn
* fix(sandbox): preset env allowances are exact names, not globs
Glob-vs-glob intersection is a losing game: every dialect extension
(character classes today) silently reopens the deny-bypass hole.
Restricting preset environment allowances to literal variable names
makes the authored-deny precedence check exact by construction: each
deny pattern is evaluated against the concrete name with the same
matcher ScrubEnv uses at runtime, so the decision cannot diverge from
enforcement regardless of deny dialect. Removes the glob intersection
machinery. Profile and --sandbox-allow env globs are unaffected.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PETCfE4crLcmodosz12qRn
* fix(sandbox): reject mandatory-deny targets in preset paths
Preset validation relied on IsSensitiveProjectTarget, which covers fewer
files than util.DANGEROUS_FILES. A preset naming .git-credentials,
.pgpass, .docker/config.json or .config/gh exactly would exact-match
suppress the mandatory deny; .git/config in allow_write would suppress
the write protection. Preset paths are now checked against
DANGEROUS_FILES (single source of truth), .git/hooks is rejected in any
direction, and .git/config is rejected for write/exec while read stays
allowed for git repo discovery. Docs state the two deliberate opt-outs
precisely.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PETCfE4crLcmodosz12qRn
* feat(sandbox): preset init and edit commands for community authoring
pmg sandbox preset init scaffolds a valid user preset (metadata flags,
threat-note template, starter rule) and refuses built-in names since
builtins win resolution. pmg sandbox preset edit opens the file via the
shared editor package and validates the result, warning when a user
preset is shadowed by a built-in. Docs lead with the scaffolded flow and
spell out builtin-vs-community provenance in preset list.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PETCfE4crLcmodosz12qRn
* refactor(sandbox): move mandatory-target matching into util
Preset path validation re-encoded knowledge util already owns: the
dangerous-files comparison and hardcoded .git/config and .git/hooks
strings. util now exports GitConfigPath, GitHooksPath (also used by
GetMandatoryDenyPatterns), PathCoveredBy and DangerousFileMatch, and
preset validation consumes them so the mandatory deny policy has a
single definition.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PETCfE4crLcmodosz12qRn
---------
Co-authored-by: Claude <noreply@anthropic.com>
* feat(sandbox): ExecutionContext plumbing and fail-closed lockdown contract
Network lockdown needs the PMG proxy's address, which is only known at
spawn time. Thread an ExecutionContext from the proxy flow through the
runner and executor into every sandbox driver, and enforce the
network_via_proxy_only fail-closed contract: lockdown without a running
loopback proxy, or on a driver that cannot enforce it, is a hard error —
never a silent fallback to unrestricted network.
- sandbox.ExecutionContext{ProxyAddr} + 4-arg Sandbox.Execute
- sandbox.ValidateLockdown validates the proxy address (loopback only)
with usefulerror code SandboxRequiresProxy
- Seatbelt validates lockdown before translation (translation itself
lands next); bubblewrap and landlock reject lockdown as unsupported
until Linux enforcement is implemented
- executor.WithExecutionContext, runner.ExecuteOptions.SandboxProxyAddr,
proxy flow passes the live proxy address
- ApplySandbox also validates centrally before invoking the driver
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PqMU5GNBbQvQct9nxek1VS
* fix(sandbox): require numeric in-range proxy port in ValidateLockdown
The validated port string is embedded into generated sandbox profiles,
so service names, zero, and out-of-range ports are refused.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PqMU5GNBbQvQct9nxek1VS
* fix(sandbox): fail closed on Seatbelt lockdown until translation lands
A lockdown policy that passed proxy validation would silently receive
the pre-lockdown network rules from the translator. Reject it until the
lockdown profile translation is implemented, keeping the window between
plumbing and enforcement fail-closed.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PqMU5GNBbQvQct9nxek1VS
* chore: review feedback — drop redundant comment, simplify stub help text
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PqMU5GNBbQvQct9nxek1VS
---------
Co-authored-by: Claude <noreply@anthropic.com>
* feat: Add support for bubblewrap sandbox
* fix: Glob pattern expansion limit for linux
* fix: Bug in glob pattern expansion for bwrap
* fix: README on trust
* fix: Multiple bubblewrap translator fix
* test: Add E2E for linux sandbox
* fix: Refactor bwrap sandbox to use common dangerous files
* fix: Path test case
* fix: Non-existent path handling bug
* refactor: Misc cleanup
* fix: Avoid bind mount for non-existentent deny protection
* fix: Off by one bug in path depth handling
* ci: Disable AppArmor on GHA runner
* fix: Disable apparmor userns restrictions
