picoclaw/pkg/tools/cron.go
nuestraai ac6e77d7c9
CLI overrides, workspace config, MagicForm channel, and security hardening (#1)
* feat(cli): add workspace, config-dir, tools, and skills override flags

Add --workspace, --config-dir, --tools, and --skills flags to
`picoclaw agent` for single-shot invocations. Supports workspace
override with bootstrap file injection (AGENTS.md, IDENTITY.md,
SOUL.md, USER.md), tool allowlisting, and skills filtering.

Also extracts FormatSkillsSummary from SkillsLoader for reuse,
adds SetSkillsFilter to ContextBuilder, and wires SkillsFilter
from agent config into the context builder.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* feat(channels): add MagicForm webhook channel plugin

Implement MagicForm channel for webhook-based agentic task delegation.
Supports Bearer token auth, async processing with HTTP callback,
workspace path validation against configurable workspace_root,
bootstrap file injection, tool/skill filtering via metadata, and
per-conversation session key isolation.

Includes Bus() accessor on BaseChannel for direct message publishing,
MagicFormConfig with workspace_root for path traversal prevention,
request body size limits, and TTL cleanup for stale request contexts.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* feat(agent): support per-request workspace, tool, and skill overrides

Add workspace override, tool allowlisting, and skills filtering to the
agent loop via processOptions metadata. When a workspace override is
active (e.g. from MagicForm channel), creates isolated SessionManager
and ContextBuilder instances per request.

Threads effSessions/effContextBuilder through all downstream paths
including runLLMIteration, forceCompression, maybeSummarize, and
summarizeSession to ensure full workspace isolation. Adds defense-in-
depth tool execution guard alongside LLM-facing tool definition filter.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* security(auth): add response body size limits to OAuth HTTP calls

Wrap all io.ReadAll(resp.Body) calls in the OAuth flow with
io.LimitReader capped at 1 MB to prevent memory exhaustion from
malicious or unexpectedly large server responses.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* security(tools): add response body size limits to search providers

Add 2 MB LimitReader to BraveSearch, Tavily, DuckDuckGo, and
Perplexity search providers to prevent memory exhaustion from
unexpectedly large responses. Matches existing pattern used by
GLMSearchProvider.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* security(tools): add 20 MB file write size limit

Reject writes exceeding 20 MB in WriteFileTool to prevent disk
exhaustion from unexpectedly large LLM-generated file content.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* security(tools): use crypto/rand for temp file naming

Replace predictable PID+nanosecond temp file names with
cryptographically random hex strings to prevent symlink
pre-placement attacks.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* security(skills): validate GitHub repo format and limit download size

Add regex validation for 'owner/repo' format to prevent URL injection
in InstallFromGitHub. Add 5 MB LimitReader on skill file downloads
to prevent memory exhaustion.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* security(auth): allow OAuth credentials to be overridden via env vars

Add PICOCLAW_OPENAI_CLIENT_ID, PICOCLAW_GOOGLE_CLIENT_ID, and
PICOCLAW_GOOGLE_CLIENT_SECRET env var overrides. Hardcoded values
remain as fallback defaults for backward compatibility.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* security(auth): sanitize OAuth error messages

Replace raw HTTP response bodies in error messages with generic
status-code-only messages. Full response details are logged at
debug level for troubleshooting.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* security(launcher): HTML-escape error messages in auth callback

Use html.EscapeString on all error text rendered in HTML responses
to prevent XSS via crafted OAuth error parameters or error messages.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* security(launcher): remove internal details from HTTP error responses

Log detailed errors server-side and return generic messages to
clients to prevent information disclosure of file paths and
internal state.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* security(launcher): add HTTP security headers middleware

Add SecurityHeaders middleware setting X-Content-Type-Options,
X-Frame-Options, and Content-Security-Policy on all responses
to mitigate MIME sniffing, clickjacking, and content injection.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* security(skills): validate ClawHub registry base URL scheme

Reject non-HTTPS registry URLs (unless localhost) to prevent
config-based redirection to malicious skill servers. Falls back
to the default https://clawhub.ai with a warning.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* security(skills): prevent workspace skills from shadowing builtins

Builtin skill names are now reserved — workspace and global skills
with the same name as a builtin are skipped with a warning log.
LoadSkill also checks builtins first. This prevents supply chain
attacks where a malicious skill replaces a trusted builtin.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* security(skills): warn when safety metadata is unavailable

Add MetadataAvailable flag to InstallResult. When the registry
metadata fetch fails (silent fallback), the user now sees a
warning that safety checks could not be completed.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* security(shell): add opt-in environment variable filtering

When filter_env is enabled in config, shell commands run with a minimal
allowlist of environment variables (PATH, HOME, LANG, TERM, etc. plus
PICOCLAW_* prefix), preventing accidental leakage of sensitive env vars
like API keys to spawned processes. Disabled by default for backward compat.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* feat(config): add workspace-local config.json overlay for per-tenant isolation

Support per-workspace config overrides via --config-dir (CLI) or configDir
(webhook), enabling different API keys, models, and agent settings per tenant.

- Add LoadWorkspaceConfig/MergeWorkspaceConfig with raw JSON overlay to
  preserve unmentioned bool fields (prevents clobbering tool enabled flags)
- Add Config.Clone() for safe per-request config copies in gateway mode
- Gateway: per-request provider creation from workspace config, with
  effProvider/effModel threaded through all LLM call sites and summarization
- CLI: workspace config merged before provider creation, CLI flags win
- MagicForm: replace inline bootstrap fields with configDir path;
  agent loop copies bootstrap files and loads config.json from configDir
- Fix --session flag: format as agent:main:cli:{key} so router honors it
- Fix cron session key: format as agent:main:cron:{id} for isolation
- Add shared CopyBootstrapFiles helper in pkg/agent/bootstrap.go
- Add config/workspace.config.example.json

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* fix(agent): honor workspace config overrides in model routing

selectCandidates uses agent.Model/Candidates baked in at startup, which
ignores per-request workspace config overrides (effProvider/effModel).
When the workspace overrides the provider, skip routing and fallback
candidates entirely to avoid cross-provider credential issues. When only
the model is overridden, allow routing but use the effective model.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* style: fix linter errors (gci, gofumpt, golines, govet, unused)

- gci: alphabetize magicform import in gateway/helpers.go
- gofumpt: reformat multi-line function signature in agent/helpers.go
- golines: wrap lines exceeding 120 chars across 5 files
- govet: rename shadowed variables (err → wcErr, s → trimmed)
- unused: remove maybeSummarize, forceCompression, summarizeSession
  wrappers superseded by their *With variants

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

---------

Co-authored-by: nuestraai <nuestraai@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Co-authored-by: admin-mf <admin@magicform.ai>
2026-03-06 10:16:36 -06:00

335 lines
10 KiB
Go

package tools
import (
"context"
"fmt"
"strings"
"time"
"github.com/sipeed/picoclaw/pkg/bus"
"github.com/sipeed/picoclaw/pkg/config"
"github.com/sipeed/picoclaw/pkg/cron"
"github.com/sipeed/picoclaw/pkg/utils"
)
// JobExecutor is the interface for executing cron jobs through the agent
type JobExecutor interface {
ProcessDirectWithChannel(ctx context.Context, content, sessionKey, channel, chatID string) (string, error)
}
// CronTool provides scheduling capabilities for the agent
type CronTool struct {
cronService *cron.CronService
executor JobExecutor
msgBus *bus.MessageBus
execTool *ExecTool
}
// NewCronTool creates a new CronTool
// execTimeout: 0 means no timeout, >0 sets the timeout duration
func NewCronTool(
cronService *cron.CronService, executor JobExecutor, msgBus *bus.MessageBus, workspace string, restrict bool,
execTimeout time.Duration, config *config.Config,
) (*CronTool, error) {
execTool, err := NewExecToolWithConfig(workspace, restrict, config)
if err != nil {
return nil, fmt.Errorf("unable to configure exec tool: %w", err)
}
execTool.SetTimeout(execTimeout)
return &CronTool{
cronService: cronService,
executor: executor,
msgBus: msgBus,
execTool: execTool,
}, nil
}
// Name returns the tool name
func (t *CronTool) Name() string {
return "cron"
}
// Description returns the tool description
func (t *CronTool) Description() string {
return "Schedule reminders, tasks, or system commands. IMPORTANT: When user asks to be reminded or scheduled, you MUST call this tool. Use 'at_seconds' for one-time reminders (e.g., 'remind me in 10 minutes' → at_seconds=600). Use 'every_seconds' ONLY for recurring tasks (e.g., 'every 2 hours' → every_seconds=7200). Use 'cron_expr' for complex recurring schedules. Use 'command' to execute shell commands directly."
}
// Parameters returns the tool parameters schema
func (t *CronTool) Parameters() map[string]any {
return map[string]any{
"type": "object",
"properties": map[string]any{
"action": map[string]any{
"type": "string",
"enum": []string{"add", "list", "remove", "enable", "disable"},
"description": "Action to perform. Use 'add' when user wants to schedule a reminder or task.",
},
"message": map[string]any{
"type": "string",
"description": "The reminder/task message to display when triggered. If 'command' is used, this describes what the command does.",
},
"command": map[string]any{
"type": "string",
"description": "Optional: Shell command to execute directly (e.g., 'df -h'). If set, the agent will run this command and report output instead of just showing the message. 'deliver' will be forced to false for commands.",
},
"at_seconds": map[string]any{
"type": "integer",
"description": "One-time reminder: seconds from now when to trigger (e.g., 600 for 10 minutes later). Use this for one-time reminders like 'remind me in 10 minutes'.",
},
"every_seconds": map[string]any{
"type": "integer",
"description": "Recurring interval in seconds (e.g., 3600 for every hour). Use this ONLY for recurring tasks like 'every 2 hours' or 'daily reminder'.",
},
"cron_expr": map[string]any{
"type": "string",
"description": "Cron expression for complex recurring schedules (e.g., '0 9 * * *' for daily at 9am). Use this for complex recurring schedules.",
},
"job_id": map[string]any{
"type": "string",
"description": "Job ID (for remove/enable/disable)",
},
"deliver": map[string]any{
"type": "boolean",
"description": "If true, send message directly to channel. If false, let agent process message (for complex tasks). Default: true",
},
},
"required": []string{"action"},
}
}
// Execute runs the tool with the given arguments
func (t *CronTool) Execute(ctx context.Context, args map[string]any) *ToolResult {
action, ok := args["action"].(string)
if !ok {
return ErrorResult("action is required")
}
switch action {
case "add":
return t.addJob(ctx, args)
case "list":
return t.listJobs()
case "remove":
return t.removeJob(args)
case "enable":
return t.enableJob(args, true)
case "disable":
return t.enableJob(args, false)
default:
return ErrorResult(fmt.Sprintf("unknown action: %s", action))
}
}
func (t *CronTool) addJob(ctx context.Context, args map[string]any) *ToolResult {
channel := ToolChannel(ctx)
chatID := ToolChatID(ctx)
if channel == "" || chatID == "" {
return ErrorResult("no session context (channel/chat_id not set). Use this tool in an active conversation.")
}
message, ok := args["message"].(string)
if !ok || message == "" {
return ErrorResult("message is required for add")
}
var schedule cron.CronSchedule
// Check for at_seconds (one-time), every_seconds (recurring), or cron_expr
atSeconds, hasAt := args["at_seconds"].(float64)
everySeconds, hasEvery := args["every_seconds"].(float64)
cronExpr, hasCron := args["cron_expr"].(string)
// Fix: type assertions return true for zero values, need additional validity checks
// This prevents LLMs that fill unused optional parameters with defaults (0) from triggering wrong type
hasAt = hasAt && atSeconds > 0
hasEvery = hasEvery && everySeconds > 0
hasCron = hasCron && cronExpr != ""
// Priority: at_seconds > every_seconds > cron_expr
if hasAt {
atMS := time.Now().UnixMilli() + int64(atSeconds)*1000
schedule = cron.CronSchedule{
Kind: "at",
AtMS: &atMS,
}
} else if hasEvery {
everyMS := int64(everySeconds) * 1000
schedule = cron.CronSchedule{
Kind: "every",
EveryMS: &everyMS,
}
} else if hasCron {
schedule = cron.CronSchedule{
Kind: "cron",
Expr: cronExpr,
}
} else {
return ErrorResult("one of at_seconds, every_seconds, or cron_expr is required")
}
// Read deliver parameter, default to true
deliver := true
if d, ok := args["deliver"].(bool); ok {
deliver = d
}
command, _ := args["command"].(string)
if command != "" {
// Commands must be processed by agent/exec tool, so deliver must be false (or handled specifically)
// Actually, let's keep deliver=false to let the system know it's not a simple chat message
// But for our new logic in ExecuteJob, we can handle it regardless of deliver flag if Payload.Command is set.
// However, logically, it's not "delivered" to chat directly as is.
deliver = false
}
// Truncate message for job name (max 30 chars)
messagePreview := utils.Truncate(message, 30)
job, err := t.cronService.AddJob(
messagePreview,
schedule,
message,
deliver,
channel,
chatID,
)
if err != nil {
return ErrorResult(fmt.Sprintf("Error adding job: %v", err))
}
if command != "" {
job.Payload.Command = command
// Need to save the updated payload
t.cronService.UpdateJob(job)
}
return SilentResult(fmt.Sprintf("Cron job added: %s (id: %s)", job.Name, job.ID))
}
func (t *CronTool) listJobs() *ToolResult {
jobs := t.cronService.ListJobs(false)
if len(jobs) == 0 {
return SilentResult("No scheduled jobs")
}
var result strings.Builder
result.WriteString("Scheduled jobs:\n")
for _, j := range jobs {
var scheduleInfo string
if j.Schedule.Kind == "every" && j.Schedule.EveryMS != nil {
scheduleInfo = fmt.Sprintf("every %ds", *j.Schedule.EveryMS/1000)
} else if j.Schedule.Kind == "cron" {
scheduleInfo = j.Schedule.Expr
} else if j.Schedule.Kind == "at" {
scheduleInfo = "one-time"
} else {
scheduleInfo = "unknown"
}
result.WriteString(fmt.Sprintf("- %s (id: %s, %s)\n", j.Name, j.ID, scheduleInfo))
}
return SilentResult(result.String())
}
func (t *CronTool) removeJob(args map[string]any) *ToolResult {
jobID, ok := args["job_id"].(string)
if !ok || jobID == "" {
return ErrorResult("job_id is required for remove")
}
if t.cronService.RemoveJob(jobID) {
return SilentResult(fmt.Sprintf("Cron job removed: %s", jobID))
}
return ErrorResult(fmt.Sprintf("Job %s not found", jobID))
}
func (t *CronTool) enableJob(args map[string]any, enable bool) *ToolResult {
jobID, ok := args["job_id"].(string)
if !ok || jobID == "" {
return ErrorResult("job_id is required for enable/disable")
}
job := t.cronService.EnableJob(jobID, enable)
if job == nil {
return ErrorResult(fmt.Sprintf("Job %s not found", jobID))
}
status := "enabled"
if !enable {
status = "disabled"
}
return SilentResult(fmt.Sprintf("Cron job '%s' %s", job.Name, status))
}
// ExecuteJob executes a cron job through the agent
func (t *CronTool) ExecuteJob(ctx context.Context, job *cron.CronJob) string {
// Get channel/chatID from job payload
channel := job.Payload.Channel
chatID := job.Payload.To
// Default values if not set
if channel == "" {
channel = "cli"
}
if chatID == "" {
chatID = "direct"
}
// Execute command if present
if job.Payload.Command != "" {
args := map[string]any{
"command": job.Payload.Command,
}
result := t.execTool.Execute(ctx, args)
var output string
if result.IsError {
output = fmt.Sprintf("Error executing scheduled command: %s", result.ForLLM)
} else {
output = fmt.Sprintf("Scheduled command '%s' executed:\n%s", job.Payload.Command, result.ForLLM)
}
pubCtx, pubCancel := context.WithTimeout(context.Background(), 5*time.Second)
defer pubCancel()
t.msgBus.PublishOutbound(pubCtx, bus.OutboundMessage{
Channel: channel,
ChatID: chatID,
Content: output,
})
return "ok"
}
// If deliver=true, send message directly without agent processing
if job.Payload.Deliver {
pubCtx, pubCancel := context.WithTimeout(context.Background(), 5*time.Second)
defer pubCancel()
t.msgBus.PublishOutbound(pubCtx, bus.OutboundMessage{
Channel: channel,
ChatID: chatID,
Content: job.Payload.Message,
})
return "ok"
}
// For deliver=false, process through agent (for complex tasks)
sessionKey := fmt.Sprintf("agent:main:cron:%s", job.ID)
// Call agent with job's message
response, err := t.executor.ProcessDirectWithChannel(
ctx,
job.Payload.Message,
sessionKey,
channel,
chatID,
)
if err != nil {
return fmt.Sprintf("Error: %v", err)
}
// Response is automatically sent via MessageBus by AgentLoop
_ = response // Will be sent by AgentLoop
return "ok"
}