* feat: add stop/backtrack capture classifications for milestone regression (#3487)
Adds 4-layer methodology for halting auto-mode and backtracking to
previous milestones when captures indicate the user wants to stop or
that a milestone missed critical features:
1. Type layer: "stop" and "backtrack" classification types in captures.ts
2. Guard layer: pre-dispatch stop check in runGuards() pauses auto-mode
before the next unit dispatches
3. Resolution layer: executeBacktrack() writes BACKTRACK-TRIGGER.md and
milestone regression markers for state machine detection
4. Protection layer: revertExecutorResolvedCaptures() detects and reverts
captures silenced by non-triage agents (resolved without classification)
Also adds fast-path stop detection in auto-post-unit.ts that pattern-matches
pending capture text for stop keywords without waiting for triage.
Closes#3487
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* feat: add slice-level skip with gsd_skip_slice tool (#3477)
Adds "skipped" as a closed status alongside "complete" and "done":
- status-guards.ts: isClosedStatus() recognizes "skipped"
- state.ts: isStatusDone() recognizes "skipped"
- gsd-db.ts: getActiveSliceFromDb() skips slices with status "skipped"
- db-tools.ts: new gsd_skip_slice tool for rethink and manual use
- rethink.md: added "Skip a slice" operation to rethink prompt
- rethink.ts: buildRethinkData shows skipped slice counts
Skipped slices satisfy dependencies for downstream slices, allowing
auto-mode to advance past them. Slice data is preserved for reference.
Relates to #3477
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: resolve 4 issues found in adversarial review of PR #3488
1. triage-ui.ts: Restore stop/backtrack entries in CLASSIFICATION_LABELS
and ALL_CLASSIFICATIONS — the Record<Classification, ...> type requires
all union members, and runtime lookups would crash on stop/backtrack.
Also auto-confirm stop/backtrack in the triage confirmation flow
(matching the triage-captures.md prompt directive).
2. triage-resolution.ts: Replace require("node:fs") in clearBacktrackTrigger
with ESM import of unlinkSync — consistent with the rest of the codebase.
3. auto-post-unit.ts: Anchor STOP_PATTERN regex to start-of-string (^) to
prevent false positives on captures like "add a pause button" or "stop
the timer from re-rendering" which are feature descriptions, not halt
directives.
4. status-guards.test.ts: Add missing test case for isClosedStatus("skipped")
to cover the new status value.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: update tool-naming test count for gsd_skip_slice
The new gsd_skip_slice tool (no alias) brings the total from 29 to 30.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
* docs: add context optimization design spec, implementation plan, and pi-layer research
- Spec: 6-change design for GSD extension context optimization
- Plan: 9-task TDD implementation plan with exact file paths and code
- Pi-layer doc: 10 infrastructure opportunities (research only, not planned)
Part of #3171, #3406, #3452, #3433.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* feat(context): add observation masking for auto-mode sessions
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* feat(context): add phase handoff anchors for auto-mode
Introduces PhaseAnchor read/write utilities so downstream agents can
inherit decisions, blockers, and intent written at phase boundaries
without re-inferring from conversation history.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* feat(context): add capability-aware model routing and context management preferences
Implement ADR-004 Phase 2 capability scoring with 7-dimension model
profiles, task requirement vectors, and weighted scoring. Add
ContextManagementConfig preferences for observation masking thresholds.
Wire capability scoring into auto-model-selection dispatch path.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* feat(context): wire observation masking, phase anchors, and tool truncation
Register observation masker in before_provider_request hook to replace
old tool results with placeholders during auto-mode. Add tool result
truncation (configurable via context_management.tool_result_max_chars).
Inject phase handoff anchors into prompt builders so downstream phases
inherit decisions from research/planning. Write anchors after successful
phase completion. Update ADR-004 status to Implemented.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* chore: remove internal planning artifacts from PR
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* docs: add capability routing, observation masking, and context management
Update dynamic-model-routing.md with capability-aware scoring section.
Update token-optimization.md with observation masking, tool truncation,
and phase handoff anchor documentation. Update configuration.md with
context_management preference block and capability_routing flag.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* Merge branch 'main' into feat/gsd-context-optimization
* fix: add context_management to known keys and prevent tool truncation state corruption
- Add missing 'context_management' to KNOWN_PREFERENCE_KEYS set so users
don't get spurious unknown-key warnings when configuring it.
- Replace in-place mutation of tool result content with immutable spread
to prevent corrupting shared conversation message objects.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: add stop and backtrack to triage-ui classification labels
The Classification type gained stop and backtrack variants from main
but triage-ui.ts was not updated, causing a TypeScript build failure.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: context masker and tool truncation operate on correct pi-ai message format
The observation masker and tool result truncation in before_provider_request
were checking m.type === "toolResult" but the actual pi-ai payload uses
m.role === "toolResult" with content as TextContent[] arrays (not strings).
bashExecution messages are converted to {role:"user"} by convertToLlm before
the hook fires, so checking m.type === "bashExecution" was a no-op.
- Fix context-masker to match on role, handle array content, detect bash
results by their "Ran `" prefix
- Fix register-hooks truncation to operate on role:"toolResult" with
array content blocks
- Update tests to use correct pi-ai LLM payload format
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
The TUI slash dispatcher started treating any unrecognized /command as handled before session.prompt() could resolve extension commands, prompt templates, or /skill:* inputs. That blocked valid non-builtin slash commands and also let /export swallow unrelated /export* prefixes.
Move unknown-command detection to the interactive entry points, allow only known builtins or session-resolved slash commands through, gate /skill:* on the skill-command setting, and tighten /export matching to exact command tokens.
Move regression tests and override tests from standalone files into
the existing test files introduced by PR #666:
- resolve-config-value.test.ts: add REGRESSION #666 describe block
and setAllowedCommandPrefixes override tests
- url-utils.test.ts: add REGRESSION #666 describe block and
setFetchAllowedUrls override tests
- Delete: regression-666.test.ts, resolve-config-value-override.test.ts,
url-utils-override.test.ts
Same 59 tests, fewer files, tests live next to the code they test.
Two regression tests that prove the bug introduced by PR #666:
1. Non-default credential tool (sops) is silently blocked by the
hardcoded SAFE_COMMAND_PREFIXES with no way to override.
2. Private IP URL is silently blocked by isBlockedUrl() with no
way to allowlist.
Both tests use dynamic import to check for the override functions,
so they run cleanly on both main (where they fail) and this branch
(where they pass). Verified in a git worktree of main.
- custom-models.md: add Command Allowlist section under Value Resolution
explaining the restriction, default list, and how to override via
allowedCommandPrefixes setting or GSD_ALLOWED_COMMAND_PREFIXES env var
- configuration.md: add URL Blocking (fetch_page) section documenting
what's blocked by default, why, and how to allowlist specific hosts
via fetchAllowedUrls setting or GSD_FETCH_ALLOWED_URLS env var
- configuration.md: add both env vars to the Environment Variables table
PR #666 introduced hardcoded SAFE_COMMAND_PREFIXES and SSRF URL
blocklists with no override mechanism. Users with non-standard
credential tools (sops, doppler, age, infisical) or needing to fetch
from internal URLs (self-hosted docs, VPN services) were silently
blocked with no recourse.
Add two global-only settings (ignored in project-level settings.json
to preserve the security property against malicious repos):
- allowedCommandPrefixes: replaces the built-in command allowlist
- fetchAllowedUrls: exempts hostnames from SSRF blocking
Both also support env var overrides (GSD_ALLOWED_COMMAND_PREFIXES,
GSD_FETCH_ALLOWED_URLS) for CI/container environments. Env vars
take precedence over settings.json.
Security model: global-only keys are stripped from project settings
at load time via stripGlobalOnlyKeys(), applied at all three
assignment points for this.projectSettings. The merge function
stays untouched — no future caller can accidentally skip stripping.
15 new tests covering override behavior, cache invalidation,
allowlist exemptions, and global-only enforcement.
Keep catch-all STREAM_RE from PR; upstream's 5-variant whack-a-mole is
superseded by the /in JSON at position \d+/ pattern. Also drop the now-
stale comment about checking stream before server/connection (no longer
needed since catch-all avoids those false-positive overlaps).
Satisfies CI require-tests gate by adding a test that verifies the
comprehensive pre-merge cleanup (step 7b) removes stale SQUASH_MSG and
MERGE_MSG files — the enhancement over the prior MERGE_HEAD-only cleanup.
https://claude.ai/code/session_01SSHD9RNwVGNxAJZEgNZpgZ
Main already had a simpler step 7c (removing only MERGE_HEAD). The PR's
step 7b is more thorough: it also removes SQUASH_MSG and MERGE_MSG,
matching the existing post-merge cleanup pattern. Replace 7c with 7b.
https://claude.ai/code/session_01SSHD9RNwVGNxAJZEgNZpgZ
Self-contained extension at src/resources/extensions/ollama/ that
auto-detects a running Ollama instance, discovers locally pulled models,
and registers them as a first-class provider with zero configuration.
Features:
- Auto-discovery of local models via /api/tags on session_start
- Capability detection (vision, reasoning, context window) for 40+ model families
- /ollama slash command with status, list, pull, remove, ps subcommands
- ollama_manage LLM-callable tool for agent-driven model operations
- Onboarding flow with auto-detect (no API key required)
- Non-blocking async probe — doesn't delay TUI paint
- Respects OLLAMA_HOST env var for non-default endpoints
Core changes (minimal):
- Add "ollama" to KnownProvider in pi-ai types
- Add "ollama" key resolution in env-api-keys.ts
- Add "ollama" default model in model-resolver.ts
- Add "Ollama (Local)" to onboarding wizard with probe flow
A pre-existing MERGE_HEAD (from failed prior merge, libgit2 native path,
or external tooling) blocks git merge --squash. Remove stale merge state
files before starting the squash merge, not just after.
When DB was available but empty, deriveState skipped deriveStateFromDb
entirely, bypassing the disk→DB sync logic. Milestones created outside
the DB write path were never discovered.
- Check HEAD~1 (newest snapshot) instead of resetTarget (pre-snapshot
base) for remote ancestry. The old check false-positived when the
remote was at the pre-snapshot base but snapshots were local-only.
- Re-run smartStage() after soft reset so RUNTIME_EXCLUSION_PATHS
apply to the absorbed commit. Without this, .gsd/ state files from
snapshot commits leaked into the real commit.
Adds a safety mechanism that detects uncommitted changes idle past a
configurable threshold (default: 30 min), auto-snapshots tracked files
using `git add -u`, and cleans up snapshot commits when real work lands.
- New `stale_uncommitted_changes` doctor issue with auto-snapshot fix
- Detection in health widget (60s), pre-dispatch gate, and /gsd doctor
- `nativeAddTracked()` stages only tracked files (no secrets/binaries)
- `absorbSnapshotCommits()` squashes `gsd snapshot:` commits into next
real autoCommit via soft reset + re-commit
- Configurable via `stale_commit_threshold_minutes` preference (0=off)
- Add extension-manifest.ts and extension-sort.ts to pi-coding-agent
with manifest reading and Kahn's BFS topological sort algorithm
- Add extensionPathsTransform hook to DefaultResourceLoader that runs
between path merging and loadExtensions() — enables pre-load
filtering and reordering without modifying pi internals
- Wire GSD's buildResourceLoader() to provide a transform that:
1. Filters ALL extensions (including community) through the GSD registry
2. Sorts in topological dependency order via sortExtensionPaths()
- Mark discoverAndLoadExtensions() as @deprecated (dead code path)
- Add 16 tests covering manifest reading, dependency sorting, cycles,
missing deps, and non-array deps
Previously, dependencies.extensions in manifests was decorative (sort
existed but was never called), and gsd extensions disable only worked
for bundled extensions. Community extensions in ~/.gsd/agent/extensions/
bypassed the registry entirely.
- Health widget: always-on last commit with relative time + message
- Dashboard: move worktree/branch info to right-aligned line under header
- Dashboard: move last commit to bottom-left with hints on right
- Dashboard: cap task titles at 45 chars, commit messages at 65 chars
- Dashboard: use … instead of ... for all truncation
70 tests covering all 16 phases of the GSD state machine with both
happy-path and failure-mode verification. Exercises DB and filesystem
derivation paths, reconciliation logic, and edge cases.
Findings documented in #3276: 0-byte SUMMARY triggers false completion,
DB task rows missing causes wrong phase, stale path cache across
derivations, non-standard status strings silently accepted.
* refactor(state): centralize pipeline logging through workflow logger
Route 15 raw process.stderr.write calls through the structured
workflow logger (logWarning/logError). Adds "db" and "dispatch"
as new LogComponent values. Enables auto-loop drain/summarize,
audit-log persistence, and doctor integration for reconciliation
and DB events that previously bypassed structured logging.
Files changed:
- workflow-logger.ts: add "db" and "dispatch" components
- state.ts: 3 reconciliation calls → logWarning/logError
- gsd-db.ts: 4 DB operation calls → logError
- workflow-reconcile.ts: 3 event merge calls → logWarning/logError
- auto-dispatch.ts: 1 reactive dispatch call → logError
- auto-post-unit.ts: 3 triage/rogue calls → logWarning/logError
* test(workflow-logger): add tests for db and dispatch log components
Cover the new LogComponent values added in this refactor to satisfy
the CI require-tests gate.
* feat(model-routing): enable dynamic routing by default
Change defaultRoutingConfig().enabled from false to true so that
dynamic model routing (tier-based downgrading for light/standard
tasks) is active out of the box. Users can still disable it via
dynamic_routing.enabled: false in PREFERENCES.md.
This is a behavioral change: sessions that previously used the
configured model for all tasks will now automatically downgrade
to cheaper models for light and standard complexity tasks.
* test(model-routing): verify dynamic routing enabled by default
Tests that defaultRoutingConfig returns enabled: true and all
routing features are active.
The catch block in reconcileMergeState silently swallowed all nativeCommit
exceptions, including real failures (permissions, corrupt git state, hook
rejections). This caused auto-mode to report success and return true (dirty,
re-derive) even when the merge commit actually failed, leading to an infinite
loop where auto-mode repeatedly attempted worktree finalization.
Now the catch block logs the error via ctx.ui.notify at "error" level and
returns false to signal that reconciliation failed, allowing upstream logic
to react appropriately. The nativeCommit return value is also checked —
a null return (nothing to commit) gets its own info notification distinct
from a successful commit SHA.
Closes#2542
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
When GSD_MILESTONE_LOCK is set (parallel worker mode), smartStage() now
excludes .gsd/milestones/<M>/ directories for all milestones other than the
locked one. This prevents a parallel worker (e.g., M033) from staging and
committing fabricated artifacts for a milestone it does not own (e.g., M032).
Previously, smartStage() ran `git add -A` with only runtime path exclusions,
allowing cross-milestone pollution when workers share the same .gsd/ directory
(git.isolation: "none"). The GSD_MILESTONE_LOCK env var only filtered what
deriveState() sees but did not prevent file staging.
Closes#1991
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
On Windows, child_process.spawn() and execFile() open a visible console
window by default. The web server spawn, RPC bridge, browser opener, and
all 15 web service subprocess calls were missing windowsHide: true,
causing constant console window flashing when running gsd --web.
Closes#2628
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
When the LLM sends an assistant message with empty content[] and
stopReason "aborted", this is a non-fatal agent stop — not a crash.
The abort handler now checks for empty content and missing errorMessage
before deciding to pause. Empty-content aborts are routed to
resolveAgentEnd instead, breaking the stuck re-dispatch loop.
Closes#2695
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Scaffolding tools (create-next-app, cargo init, etc.) create nested .git
directories inside worktrees. Git records these as gitlinks (mode 160000)
without .gitmodules, so worktree cleanup destroys the only copy of the
nested object database — causing permanent silent data loss.
Added findNestedGitDirs() helper that recursively scans worktree for nested
.git directories (skipping node_modules and other non-project dirs). The
removeWorktree() function now calls this before cleanup and removes any
nested .git dirs so files are tracked as regular content.
Closes#2616
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
When the default isolation mode flipped from "worktree" to "none" between
versions, mergeAndExit() returned early for mode "none" without checking
whether the session was physically inside an active worktree. This silently
skipped the merge, orphaning committed work on the milestone branch.
The fix moves the worktree-presence check (isInAutoWorktree + originalBasePath)
before the mode-none early return. If we are inside a worktree, mergeAndExit
proceeds with the worktree merge path regardless of the configured mode.
Also fixes the misleading JSDoc on GitPreferences.isolation that claimed
"worktree" was the default when the runtime default is actually "none".
Closes#2625
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
When an agent requests read(file, offset: 30) on a 13-line file, the
read tool threw "Offset 30 is beyond end of file" which propagated as
invalid JSON downstream during milestone completion. Now clamps the
offset to the last line and prepends a notice, allowing the agent to
continue with valid content.
Fixes both read.ts and hashline-read.ts variants.
Closes#3007
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
isGhostMilestone() now checks for DB rows and worktree directories before
falling back to content-file detection. A milestone with a DB row or a
worktree is a legitimate milestone that hasn't been populated yet, not a
ghost from a killed session.
Fixes#2921
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
When a session grows beyond the context window of available models,
generateSummary() now detects the overflow and falls back to chunked
summarization: split messages into context-fitting chunks, summarize
the first chunk, then iteratively merge subsequent chunks using the
existing UPDATE_SUMMARIZATION_PROMPT path.
Closes#2932
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Two root causes destroyed terminal state during normal navigation:
1. The pagehide handler fired a shutdown beacon unconditionally, but on
mobile/Safari tab switches pagehide fires with event.persisted=true
(bfcache entry). This killed the server and all PTY sessions when the
user merely switched browser tabs. Fix: check event.persisted and skip
the beacon when the page is being cached, not unloaded.
2. ShellTerminal used project-agnostic session IDs ("default"), so
switching projects and switching back either collided with the old
session or spawned a new one, losing terminal state. Fix: scope session
IDs by project path (e.g. "default:/path/to/project") so the server's
getOrCreateSession returns the existing live PTY on reconnect.
Closes#2701
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
cleanupQuickBranch() was exported from quick.ts but never called anywhere.
After a /gsd quick task completed, the user was left on the quick branch
with orphaned state in quick-return.json.
Register a turn_end hook in register-hooks.ts that calls cleanupQuickBranch()
after each agent turn. The function is already idempotent (no-op when no
quick-return state is pending), so it is safe to call on every turn.
Closes#2668
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
