Sn Image Resume

sn-image-resume

Resume image generation scene skill (tier 1), relying on the sn-text-optimize and sn-image-generate tools provided by sn-image-base (tier 0).

Features:

• Accepts resume content directly from conversational text
• Supports optional user-provided style direction
• Applies the fixed portfolio-resume layout rules in prompts/resume.md
• Generates a tall designed resume image through sn-image-generate

Non-goals

• Editing or polishing a plain text resume document without generating an image
• Parsing uploaded resume files as the primary input format
• Creating a conventional single-column ATS resume
• Guaranteeing exact preservation of every long paragraph when the image layout requires compression

Input Specification

API Configuration

All API calls in this skill are executed through the sn_agent_runner.py of the sn-image-base skill, with authentication parameters using default values (CLI > environment variables > built-in defaults), so they do not need to be passed explicitly in normal use.

If all capabilities use the same gateway, configure only:

iniCopy code
SN_BASE_URL="https://your-api-endpoint.com/v1"
SN_API_KEY="your-api-key"

When encountering MissingApiKeyError or needing to specify a model: pass parameters explicitly via CLI. See $SN_IMAGE_BASE/references/api_spec.md.

$SN_IMAGE_BASE path explanation: $SN_IMAGE_BASE is the installation directory of the sn-image-base skill (SKILL.md exists). The agent can locate this path by skill name sn-image-base.

Architecture: Main Agent + Worker Agent

This skill uses a two-tier agent architecture:

Responsibility Boundaries:

• Worker Agent does not send any messages to the user directly, only returns structured JSON
• Main Agent is responsible for all user-visible messages
• Worker Agent's last message must be and only be the JSON string defined in the Return Contract
• Worker Agent's low-level API calls execute directly through sn-image-base, without spawning nested subagents

Workflow

Main Agent Workflow

1. Extract resume_content, optional style, aspect_ratio (default 9:16), image_size (default 2k), and output_mode (default friendly) from the user request
2. Validate that resume_content is non-empty and contains enough resume information to generate a meaningful page
3. Validate aspect_ratio against the allowed values: 2:3, 3:2, 3:4, 4:3, 4:5, 5:4, 1:1, 16:9, 9:16, 21:9, 9:21. If the user-provided value is not in this list, inform the user and fall back to the default 9:16
4. Send uniform preflight message: "Using sn-image-resume skill to generate a resume image, please wait..."
5. Start Worker Agent, passing in complete parameters and working directory
6. When Worker Agent returns:
   • status=ok: send a short summary and the generated image
   • status=error: report the real error field content to the user

Worker Agent Workflow

Worker Agent receives resume_content, style, aspect_ratio, image_size, output_mode, and the working directory of this skill (SKILL_DIR).

Step 0 — Initialization

1. Generate task_id using timestamp format YYYYMMDD_HHMMSS
2. Create temporary directory: /tmp/openclaw/sn-image-resume/<task_id>/ as TEMP_DIR
3. Persist normalized inputs:

bashCopy code
echo "$RESUME_CONTENT" > "$TEMP_DIR/resume-content.txt"
echo "$STYLE" > "$TEMP_DIR/style.txt"

Step 1 — Resume Prompt Generation

Use prompts/resume.md as the system prompt and call sn-text-optimize to convert the user resume content into a detailed image generation prompt.

bashCopy code
USER_PROMPT=$(cat << EOF
Resume content:
$RESUME_CONTENT

Optional style instruction:
${STYLE:-No explicit style instruction. Infer an appropriate professional visual style from the resume content.}

Task:
Convert the resume content into a complete text-to-image prompt for a tall portfolio-resume image.
Follow the fixed layout, language, content mapping, typography, panel, and style translation rules in the system prompt.
Return only the final image generation prompt. Do not include explanations, markdown fences, or alternative options.
EOF
)

python "$SN_IMAGE_BASE/scripts/sn_agent_runner.py" sn-text-optimize \
  --system-prompt-path "$SKILL_DIR/prompts/resume.md" \
  --user-prompt "$USER_PROMPT" \
  --output-format json

Parse JSON stdout and extract result as generation_prompt. If the process exits non-zero, returns invalid JSON, or result is empty, return status=error with the actual error.

Persist output:

bashCopy code
echo "$GENERATION_PROMPT" > "$TEMP_DIR/generation-prompt.txt"

Step 2 — Resume Image Generation

Generate the final image using sn-image-base's sn-image-generate tool.

bashCopy code
python "$SN_IMAGE_BASE/scripts/sn_agent_runner.py" sn-image-generate \
  --prompt "$GENERATION_PROMPT" \
  --image-size "$IMAGE_SIZE" \
  --aspect-ratio "$ASPECT_RATIO" \
  --save-path "$TEMP_DIR/resume.png" \
  --output-format json

Parse JSON stdout. If generation fails, return status=error with the actual error. The generated image path is $TEMP_DIR/resume.png.

Error Handling Rules

• If required resume content is missing, ask the user to provide resume text before starting generation
• If sn-text-optimize fails or returns an empty result, stop and report the real error
• If sn-image-generate fails, stop and report the real error
• Do not silently substitute a generic resume prompt when user content is incomplete or prompt generation fails
• Do not invent factual resume details that the user did not provide; only reorganize, condense, and visually map provided information

Return Contract

After Worker Agent completes, its last message must be and only be the following JSON string (bare JSON, no code fences, no preceding or trailing text).

Normal Flow:

jsonCopy code
{
  "status": "ok",
  "need_main_agent_send": true,
  "output_mode": "friendly|verbose",
  "image": "$TEMP_DIR/resume.png",
  "generation_prompt": "<included only when output_mode=verbose>",
  "timing": {
    "total_elapsed_seconds": 25.12,
    "prompt_generation": { "elapsed_seconds": 5.23, "model": "sensenova-6.8-flash-lite" },
    "image_generation": { "elapsed_seconds": 19.89, "model": "sn_image_model" }
  }
}

Error Flow:

jsonCopy code
{
  "status": "error",
  "error": "<Actual error information>"
}

Rules:

• status=ok must contain need_main_agent_send: true
• generation_prompt must contain when output_mode=verbose; omit it in friendly mode
• timing.prompt_generation.elapsed_seconds and timing.prompt_generation.model are read from sn-text-optimize JSON output
• timing.image_generation.elapsed_seconds is read from sn-image-generate JSON output
• timing.image_generation.model is fixed to "sn_image_model" because sn-image-generate does not return a model field

Output Format

friendly mode (default)

Text Summary: one sentence describing that the resume image has been generated, no more than 50 words.

Image: send the single generated resume image.

verbose mode

textCopy code
Resume image generated
---
Aspect ratio: <aspect_ratio>
Image size: <image_size>
---
Generation prompt:
<generation_prompt>
---
Time statistics: Total <total>s | Prompt generation <t>s | Image generation <t>s
---
Image:
<image path>

Call Relationship

• Bottom-level dependency: sn-image-base → sn-text-optimize, sn-image-generate
• System prompt: prompts/resume.md

References

• prompts/resume.md - Fixed portfolio-resume layout and language/content mapping rules
• ../sn-image-base/SKILL.md - Base-layer image/text tool behavior
• ../sn-image-base/references/api_spec.md - CLI parameter details