# extensions/builtin/

> P2 | Parent: ../AGENT.md

Member List
diagnostics/index.ts: Diagnostics extension entry, subscribes to diagnostic:event, buffers session-local diagnostic records, prompts only after threshold at agent_end, registers /report-issue
goal/index.ts: Goal extension entry, per-thread GoalController, /goal command and goal subcommand autocomplete, get_goal/create_goal/update_goal tools, turn/account lifecycle hooks, idle-continuation prompt dispatch, GOAL_MESSAGE_TYPE renderer, session_start/shutdown hooks, persistent per-thread status footer indicator
goal/goal-types.ts: ThreadGoalStatus enum (active/paused/blocked/usage_limited/budget_limited/complete), ThreadGoal record, GoalSetMode, GoalAccountingMode, GoalTurnAccounting, GoalControllerState, validation constants
goal/goal-store.ts: GoalStore JSON-file persistence layer; get/replace/insert/update/delete/account_usage primitives, atomic temp-file rename, status-filtered accounting modes, budget-limit auto-downgrade
goal/goal-format.ts: formatGoalElapsedSeconds, formatTokens, goalStatusLabel, goalUsageSummary, goalSummaryLines, goalStatusIndicator, shouldConfirmBeforeReplacing, editedGoalStatus, validateObjective, validateBudget
goal/goal-prompts.ts: buildContinuationPrompt, buildBudgetLimitPrompt, buildObjectiveUpdatedPrompt - three steering templates mirroring codex-rs prompts/templates/goals/*
goal/goal-controller.ts: GoalController class, per-thread mutex, on_turn_start/on_token_usage/on_tool_finish/on_turn_end/on_turn_abort/on_turn_error/on_usage_limit hooks, idle continuation followUp injection, blocked-signal escalation counter, budget-limit steering emission
goal/goal-tools.ts: get_goal/create_goal/update_goal LLM tool factories, TypeBox parameter schemas, host singleton for controller lookup, renderGoalResponse formatter
goal/goal-parser.ts: parseGoalCommand, buildGoalHelp, getGoalArgumentCompletions for /goal subcommand autocomplete
goal/goal-command.ts: runGoalCommand handler for /goal show/clear/edit/pause/resume/set with ConfirmIfExists dialog, multi-line summary renderer
goal/README.md: Goal extension documentation - usage, LLM tools, lifecycle, persistence, status indicator, file-by-file architecture
diagnostics/types.ts: Diagnostic event/report type contract and diagnostic:event channel name
diagnostics/diagnostic-buffer.ts: DiagnosticBuffer, event coercion, fingerprint dedupe, prompt gating
diagnostics/reporter.ts: User-approved InsForge catui_issue_events reporter, configured via CATUI_ISSUE_* env vars
diagnostics/redaction.ts: Diagnostic message normalization and secret/path redaction helpers
browser/index.ts: Browser Harness extension entry, registers browser/browser_admin tools, /browser command, Browser Harness resource discovery, project-local browser workspace seeding; opt-in since P6/EV03 despite physical source staying under builtin pending Q2 physical/package decision
browser/browser.md: Browser Harness day-to-day skill instructions for Catui tool use and workspace contribution
browser/install.md: Browser Harness setup and troubleshooting instructions, exposed as a skill resource
browser/src/browser_harness/: Vendored Browser Harness Python package, CDP daemon, IPC bridge, admin commands, and helper functions
browser/interaction-skills/: Reusable Browser Harness mechanics guides for browser interactions
browser/agent-workspace/: Seed workspace copied to .catui/browser-workspace for editable helpers and domain skills
discipline/index.ts: Engineering discipline extension entry, registers skill tool, default workflow skills, and lightweight before_agent_start bootstrap prompt
discipline/skills/: Built-in engineering workflow skills for brainstorming, debugging, TDD, verification, planning, code review, worktrees, and branch finishing
idle-think/index.ts: IdleThink extension entry, session lifecycle registration, activity tracking, and persistent insight prompt injection
idle-think/idle-think-runtime.ts: IdleThink runtime boundary - state, budget reset, interval ownership, abort cleanup, exploration orchestration, insight persistence, and diagnostic reporting
idle-think/thinker.ts: IdleThink read-only SubAgent exploration runner, project context collection, network capability detection, and exploration prompt construction
idle-think/insights.ts: IdleThink nanomem-backed project-scoped insight storage and system prompt injection
idle-think/curiosity.ts: IdleThink curiosity queue persistence, topic selection, dedupe, and explored-topic pruning
link-world/index.ts: Internet access integration entry, registers link_world_admin/link_world_exec/web_search/web_fetch tools, /link-world status/doctor/version/install commands, and bundled internet-search resources
link-world/link-world-agent.md: Agent-facing skill that tells models to prefer web_search/web_fetch and link_world_admin/link_world_exec over bash for internet tasks
link-world/network-routing.md: Routing skill that tells models when to use web_search/web_fetch/link-world versus browser automation
link-world/agent-workspace/: Seed workspace copied to .catui/link-world-workspace for project-local domain skills and notes
mcp/index.ts: MCP protocol integration extension, MCP guidance resources
next-step/index.ts: Default-on Codex-style "suggest next steps" rule injected into agent systemPrompt via before_agent_start; settings.nextStep.enabled guard (default true); NEXT_STEP_RULE export for testing
presence/index.ts: AI-driven opening + idle presence lines, git/cwd snapshot, systemPrompt injection of recent presence lines, timers/debounce/idle in-flight lock, settings.presence.enabled guard, PRESENCE_MESSAGE_TYPE renderer
presence/presence-memory.ts: Presence memory boundary - NanoMem directory/project discovery, memory-derived locale detection, randomized preference/lesson highlight selection for greeting prompts
plan/index.ts: Plan Mode extension entry, /plan /plan:validate /plan:approve commands, EnterPlanMode/ExitPlanMode tools, permission gating, TUI status/widget, workflow prompt injection
plan/types.ts: Plan mode types, persisted state payload, attachment variants, permission result model
plan/plan-file-manager.ts: Plan file path management, slug generation, settings-aware plans directory, plan state hydration/serialization, resume/fork copy helpers
plan/plan-permissions.ts: Plan mode tool permission gating, exact plan-file write checks, read-only bash allowlist, agent/tool blocking policy
plan/plan-workflow-prompt.ts: Plan mode full/sparse workflow prompt, reentry prompt, exit prompt, EnterPlanMode/ExitPlanMode tool result text
plan/enter-plan-mode-tool.ts: EnterPlanMode tool for model-initiated plan mode entry and UI state activation
plan/exit-plan-mode-tool.ts: ExitPlanMode tool for user-approved plan mode exit, validation, teammate approval, UI cleanup
plan/plan-agents.ts: Explore/Plan subagent prompt helpers and read-only tool list for plan mode
plan/plan-validation.ts: Plan structure validation for Context, Approach, Files/Changes, and Verification sections
plan/teammate-approval.ts: Teammate plan approval mailbox integration and approval/waiting message formatting
subagent/index.ts: SubAgent extension entry, /subagent:/subagent:run/:stop/:status/:report/:apply commands, SUBAGENT_MESSAGE_TYPE renderer
subagent/subagent-parser.ts: SubAgent command parsing, parseSubAgentCommand/buildSubAgentHelp
subagent/subagent-runner.ts: SubAgent orchestration — research (read-only) and implement (isolated worktree) roles, diff preview and apply flow
subagent/subagent-types.ts: SubAgent extension types — SubAgentPhase, SubAgentWorkerInfo, SubAgentRunState, SubAgentRunReport
interview/index.ts: Requirement clarification extension entry, /interview and /grill-me commands, interview tool registration, custom message renderers, lightweight before_agent_start hook
interview/interview-runtime.ts: Interview probe/runtime boundary - synchronous prompt heuristics, workspace context collection, LLM probe coercion, UI follow-up loop, refined intent injection text
security-audit/interface.ts: Security audit interface, SecurityCheckResult/AuditEvent/SecurityEngine
security-audit/index.ts: Security extension entry, audit logging and dangerous pattern detection
security-audit/engine/interceptor.ts: Request/response interception, InterceptorResult confirmation flow
security-audit/engine/logger.ts: Security event logging, JSON file audit trail
security-audit/engine/detector.ts: Vulnerability detection, pattern matching for dangerous commands
soul/index.ts: AI personality evolution extension, persistent personality across sessions
grub/index.ts: Grub extension entry - autonomous iterative task runner, /grub command, dual-phase prompt injection, feature-list guard dispatch, GRUB_MESSAGE_TYPE renderer
grub/grub-controller.ts: GrubController - drives autonomous grub iterations, durable GrubTaskState, initializer baseline capture, feature-list mutation validation
grub/grub-decision.ts: Grub assistant protocol parser, extracts validated loop-state decisions from assistant text
grub/grub-parser.ts: Grub command parsing, parseGrubCommand/buildGrubHelp
grub/grub-prompts.ts: Grub prompt construction boundary, initializer/coding system prompts and per-task dispatch prompts
grub/grub-harness.ts: Grub harness artifact boundary, .grub/<id>/ feature-list/progress-log/init.sh creation
grub/grub-format.ts: Grub user-facing status/result formatter for readable TUI messages
grub/grub-turn.ts: Grub turn-end coordinator, parsing assistant output, checklist gates, retry/terminal update events
grub/grub-i18n.ts: Grub localization helper, English/Chinese prompts and TUI strings
grub/grub-feature-list.ts: feature-list.json IO, initializer skeleton, legacy checklist migration, passes/evidence-only diff validation
grub/grub-persistence.ts: Cross-session .grub/<id>/state.json persistence, active task discovery, stale harness pruning
grub/grub-types.ts: Grub types, GrubStatus/GrubDecisionStatus/GrubDecision/GrubTaskState/GrubTaskSnapshot/FeatureList/PersistedGrubState
grub/README.md: Grub extension documentation - autonomous "keep digging until done" runner without default git commits
loop/index.ts: Loop extension entry - session-scoped recurring prompt/command scheduler with pause/resume/run-now/max-runs/quiet, /loop command + LOOP_MESSAGE_TYPE renderer
loop/scheduler-controller.ts: SchedulerController - in-memory recurring task store with pause/resume/run-now/max-runs, MAX_SCHEDULED_TASKS=50
loop/scheduler-parser.ts: Loop command parsing with flags/subcommands, parseSchedulerCommand/parseDurationSpec/buildSchedulerHelp, --name/--max/--quiet
loop/scheduler-types.ts: Scheduled loop types, LoopPayloadKind/ScheduledLoopTask/LoopStartSpec/ParsedSchedulerCommand
loop/README.md: Loop extension documentation - recurring scheduler usage and flags
sal/index.ts: SAL extension entry, enabled by default, registers flags, /sal:* commands, lifecycle hooks including agent_result, terrain snapshot refresh, eval event emission, and stale-run cleanup scheduling; delegates config, context, runtime contracts, and tool_trace analytics to focused SAL modules
sal/sal-config.ts: SAL build metadata, eval environment constants, credential loading, truthy parsing, stale-cleanup/A-B flag resolution, experiment id normalization, and sidecar directory resolution
sal/sal-context.ts: SAL anchor system-prompt injection formatting plus A/B sidecar turn-record persistence
sal/sal-runtime.ts: SAL shared BuildMeta/TurnState/SalRuntime contracts used across config, context, trace, and entry modules, including per-turn loop outcome state
sal/sal-trace.ts: SAL tool path extraction, task intent inference, and bounded tool_trace payload construction with loop outcome summary
sal/terrain.ts: TerrainSnapshot/TerrainNode/TerrainEdge model, buildTerrainIndex(), checkDipCoverage(), isSnapshotStale(), moduleIdForPath(), parses P2 AGENT.md and P3 file headers
sal/anchors.ts: StructuralAnchor/AnchorResolution model, locateTask(), locateAction(), evidence-driven scoring with tunable SalWeights, CJK bigram tokenization
sal/weights.ts: SalWeights interface, SAL_DEFAULT_WEIGHTS, loadSalWeights() reads sal-config.json from workspace or .memory-experiments/sal/
sal/eval/index.ts: createEvalSink() factory + barrel re-exports; adapter selection via options.adapter or endpoint scheme inference (http(s)→insforge, file://|/|./|../→jsonl, missing→noop); ONLY entry point SAL imports from
sal/eval/types.ts: EvalSink interface, EvalEventEnvelope/EvalEventType (run_start/run_end/turn_anchor), EvalAdapterId ("insforge"|"jsonl"|"noop"), CreateEvalSinkOptions with optional onDiagnostic callback, createEvalEvent factory; zero-dependency type surface
sal/eval/noop-sink.ts: noopSink — silent EvalSink used when eval disabled or no adapter configured
sal/eval/insforge-sink.ts: InsForgeEvalSink — PostgREST adapter, routes run_start→eval_runs INSERT (merge-duplicates) with legacy-schema fallback, writes turn_anchor/tool_trace/memory_recalls/run_end only after parent run confirmation, tool_trace→eval_tool_traces with PGRST204 fallback, memory_recalls→eval_memory_recalls batch INSERT, run_end→eval_runs PATCH; allowSelfSigned TLS option logs only in development runtime, batching with default 2000ms interval
sal/eval/jsonl-sink.ts: JsonlEvalSink — append-only filesystem adapter, one JSON object per line, accepts file:// URLs or plain paths, auto-creates parent dir, batched writes
sal/README.md: SAL extension usage, sidecar output layout, weights override, pluggability contract
team/index.ts: AgentTeam extension entry, /team <task>, /team:spawn/:preset/:send/:status/:progress/:psyche/:dashboard/:task/:mail/:allow-path/:stop/:terminate/:approve/:mode command registration and dispatch
team/team-ui.ts: AgentTeam message renderer, list/status/task formatting, dashboard visibility/timer ownership, realtime observer, speaker-stream emission helpers
team/team-types.ts: TeammateRole/TeammateMode/TeammateStatus/TeamTask/HarnessState/PsycheWeights/TeamUtterance/Handoff/LeaderPlan/AgentLiveView/TeammateIdentity/PersistedTeammate/TeamSpawnSpec/TeamSendResult types
team/team-state-store.ts: TeamStateStore class - durable teammate persistence via JSON files in <agentDir>/teams/
team/team-parser.ts: Team command parser - parseTeamCommand/buildTeamHelp for /team:* subcommands
team/team-runtime.ts: TeamRuntime class - teammate registry, stable internal labels plus visible teammate names, per-teammate send queue, lifecycle, durable tasks, mailbox + permission + transcript wiring
team/team-runtime-helpers.ts: TeamRuntime helper boundary - teammate prompt construction, harness turn preparation, live event projection, tool selection, path guards, label/role/text helpers
team/team-orchestrator.ts: Leader orchestration helpers - plan building, speaker utterance creation, @mention parsing, and handoff execution
team/team-task-store.ts: TeamTaskStore class - durable shared task list in <storageDir>/tasks.json
team/team-harness.ts: Harness protocol helpers - context files, phase instructions, checkpoint/revert, feature validation
team/team-presets.ts: Preset definitions and executor - solo/duo/squad spawning, model-assisted auto team selection, heuristic fallback
team/team-dashboard.ts: Text dashboard/status rendering - teammate workbench cards, current task, last utterance, progress bars, footer summary
team/team-psyche.ts: Psyche prompt layer - role/phase weighted Id/Ego/Superego prompt construction
team/team-permissions.ts: PermissionStore - pending permission request queue, approve/deny, path allowlists
team/team-mailbox.ts: TeamMailbox - typed JSONL-backed append-only message log for leader↔teammate and teammate↔teammate
team/team-transcript.ts: TeamTranscriptWriter - per-teammate JSONL transcripts
team/TESTING.md: Manual & smoke-test guide for Phase B AgentTeam

Rule: Members complete, one item per line, parent links valid, precise terms first

[COVENANT]: Update this file header on changes and verify against parent AGENT.md