executus/tools/file_get_text.go

// file_get_text fetches a saved text file's content as plain text.
// Only succeeds for text/* MIMEs; binary MIMEs return an error so the
// agent knows to use a different path (file_get_metadata for
// reasoning, send_attachments for delivery).
//
// Why a 64 KiB cap: the v10 byte-vs-reference principle says inline
// text content stays under ~10KB ideally; we set the hard cap at 64
// KiB to handle reasonable text artifacts (logs, configs, small
// reports) without blowing the agent's context. Files larger than
// the cap return an error pointing to send_attachments.
//
// Why a separate tool (vs file_get): file_get returns base64 +
// metadata regardless of MIME, which agents misuse to dump 10MB PDFs
// into the context window. file_get_text is the agent-friendly
// alternative that explicitly fails fast on binary content.
package tools

import (
	"context"
	"encoding/json"
	"errors"
	"fmt"
	"strings"
	"time"

	"gitea.stevedudenhoeffer.com/steve/executus/tool"
)

const fileGetTextMaxBytes = 64 * 1024

type fileGetTextArgs struct {
	FileID string `json:"file_id" description:"Opaque file ID returned by file_save or file_list."`
}

type fileGetTextResult struct {
	Text      string `json:"text"`
	Mime      string `json:"mime"`
	SizeBytes int64  `json:"size_bytes"`
	CreatedAt string `json:"created_at"` // RFC3339
}

// NewFileGetText constructs the file_get_text tool. storage nil →
// "not configured" at execute time.
func NewFileGetText(storage FileStorage) tool.Tool {
	return tool.NewGatedTool[fileGetTextArgs](
		"file_get_text",
		"Fetch a saved text file's content (text/* MIMEs only, capped at 64KB). For binary content use file_get_metadata + send_attachments. Errors with 'not_text' for non-text MIMEs and 'too_large' for files > 64KB.",
		tool.Permission{
			AuthoringRequirement: tool.RequirementAnyone,
			OperatesOn:           tool.ScopeCaller,
			SafeForShare:         true,
			Categories:           []string{"storage", "read"},
		},
		func(ctx context.Context, inv tool.Invocation, args fileGetTextArgs) (string, error) {
			if storage == nil {
				return "", fmt.Errorf("file_get_text: not configured")
			}
			if args.FileID == "" {
				return "", fmt.Errorf("file_get_text: file_id required")
			}
			meta, content, err := storage.FileGet(ctx, args.FileID)
			if err != nil {
				if errors.Is(err, ErrFileNotFound) {
					return "", fmt.Errorf("file_get_text: not found")
				}
				return "", fmt.Errorf("file_get_text: %w", err)
			}
			// Descendant grant: a worker this run (transitively)
			// dispatched may have produced the file — its scope is the
			// WORKER's run, so the grant also stands in for the scope
			// check below.
			grantedViaDescendant := false
			if meta.SkillID != inv.SkillID {
				if !descendantFileGrant(ctx, storage, inv, meta.SkillID) {
					return "", fmt.Errorf("file_get_text: file does not belong to this skill")
				}
				grantedViaDescendant = true
			}
			if !grantedViaDescendant {
				if err := ValidateScope(inv, meta.Scope, false); err != nil {
					return "", fmt.Errorf("file_get_text: %w", err)
				}
			}
			if !isTextMime(meta.MimeType) {
				return "", fmt.Errorf("file_get_text: not_text: mime %q is not text/*", meta.MimeType)
			}
			if int64(len(content)) > fileGetTextMaxBytes {
				return "", fmt.Errorf("file_get_text: too_large: %d bytes exceeds 64KB cap; use send_attachments to deliver this file to Discord", len(content))
			}
			res := fileGetTextResult{
				Text:      string(content),
				Mime:      meta.MimeType,
				SizeBytes: meta.SizeBytes,
				CreatedAt: meta.CreatedAt.UTC().Format(time.RFC3339),
			}
			b, err := json.Marshal(res)
			if err != nil {
				return "", fmt.Errorf("file_get_text: marshal: %w", err)
			}
			return string(b), nil
		},
	)
}

// isTextMime reports whether the given MIME is a text/* type.
// Accepts "text/plain", "text/markdown", "text/csv", "application/json"
// and "application/xml" since those are conventionally text.
func isTextMime(mime string) bool {
	mime = strings.ToLower(strings.TrimSpace(mime))
	if strings.HasPrefix(mime, "text/") {
		return true
	}
	switch mime {
	case "application/json", "application/xml", "application/xhtml+xml",
		"application/javascript", "application/yaml", "application/x-yaml":
		return true
	}
	return false
}