reader/docs/openclaw/openclaw-handoff.md

# OpenClaw Handoff

## Purpose

This repository provides a FreshRSS-first reading pipeline for OpenClaw:

`FreshRSS unread items -> RSS content extraction -> LLM summary -> rule engine -> OpenClaw delivery payload`

OpenClaw should treat this repository as an MCP-backed upstream content processor.

This repository is responsible only for upstream reading-pipeline work:

- FreshRSS pull
- content extraction
- LLM summary generation/validation
- rule-based filtering
- OpenClaw delivery payload generation
- selected-article summary capability based on existing extracted text
- run-state persistence, status lookup, result lookup, and minimal resume for the FreshRSS workflow

This repository should **not** take over downstream orchestration responsibilities that belong to OpenClaw / skills, such as:

- Hugo publishing
- chat reporting
- user confirmation handling
- IMA upload orchestration

## Production Entrypoint

reader 当前正式工作流服务入口是 MCP tool:

- `run_freshrss_openclaw_pipeline`

It starts the only formally supported workflow today: `freshrss_daily_digest`.

OpenClaw should treat the returned `run_id` as the only stable handle for follow-up reads. Do not hand-build `outputs/freshrss/rerun/...` paths in OpenClaw.

## Supported MCP Tools

Current MCP tools: 11 total.

Workflow service tools:

- `run_freshrss_openclaw_pipeline`
- `get_run_status`
- `list_runs`
- `list_run_artifacts`
- `get_delivery_payload`
- `get_run_report`
- `resume_run`

Single-step / debug tools:

- `extract_url_content`
- `extract_item_content`
- `filter_summary_result`
- `generate_article_summaries`

## Required Environment Variables

The MCP server process must have these variables available:

- `FRESHRSS_API_BASE_URL`
- `FRESHRSS_USERNAME`
- `FRESHRSS_API_PASSWORD`
- `LLM_API_URL`
- `LLM_API_KEY`
- `LLM_MODEL`

Example:

```powershell
set FRESHRSS_API_BASE_URL=http://127.0.0.1:8081/api/greader.php
set FRESHRSS_USERNAME=osiman
set FRESHRSS_API_PASSWORD=your-api-password
set LLM_API_URL=https://api.deepseek.com
set LLM_API_KEY=your-llm-api-key
set LLM_MODEL=deepseek-chat
```

## Server Startup

Install dependencies:

```bash
pip install -e .
```

Start the MCP server:

```bash
summary-mcp
```

## Recommended MCP Workflow

Recommended production path:

1. Call `run_freshrss_openclaw_pipeline` and persist the returned `run_id`
2. Use `get_run_status(run_id)` as the authoritative run-state read for status, stage, artifacts, and recovery
3. Use `list_runs(...)` when OpenClaw needs recent-run discovery or high-level inspection
4. Use `list_run_artifacts(run_id)` when OpenClaw needs to inspect what this run actually produced
5. Use `get_delivery_payload(run_id)` and `get_run_report(run_id)` as the formal result-reading APIs
6. Use `resume_run(run_id)` only when the run falls inside the minimal supported resume scope

OpenClaw should not directly derive or hardcode:

- `outputs/freshrss/rerun/<run_dir>/run-state.json`
- `outputs/freshrss/rerun/<run_dir>/candidates/openclaw-delivery-payload.json`
- `outputs/freshrss/rerun/<run_dir>/run-report.json`

If filesystem access is needed for debugging, consume only paths returned by MCP such as `output_dir`, `artifact.path`, `delivery_output`, or `report_output`.

## Recommended MCP Call

Recommended production start call:

```json
{
  "limit": 5,
  "mark_read": true,
  "include_read": false,
  "debug_artifacts": false,
  "timeout_seconds": 60,
  "max_retries": 2
}
```

Recommended semantics:

- Use `mark_read=true` for normal production runs.
- Use `mark_read=false` only for debug, test, or validation runs.
- Keep `debug_artifacts=false` for routine production runs.
- Set `debug_artifacts=true` only when troubleshooting a bad batch.
- Treat the returned `run_id` as the stable identifier for all follow-up MCP reads.
- If no real `openclaw-delivery-payload.json` was produced, OpenClaw should stop instead of generating a digest from placeholders or examples.

## Formal Capability Boundary

reader 当前正式 MCP workflow service 的边界如下：

- formal workflow: only `freshrss_daily_digest`
- run truth: every FreshRSS run writes `run-state.json`
- state query tools: `get_run_status`, `list_runs`, `list_run_artifacts`
- result read tools: `get_delivery_payload`, `get_run_report`
- `digest-brief.json` is generated and registered as an artifact, but there is no standalone `get_digest_brief` tool yet
- `run_freshrss_openclaw_pipeline` and `resume_run` are synchronous MCP calls today; there is no background queue / worker model yet
- `generate_article_summaries` is supported, but it is outside the formal `resume_run` scope and not part of the FreshRSS workflow-state model

Historical compatibility note:

- `get_run_status` / `list_runs` / `list_run_artifacts` can still infer basic state for older run directories without `run-state.json`
- `resume_run` does **not** support those inferred historical runs; it requires a valid `run-state.json`

## What The Tool Returns

Primary return fields from `run_freshrss_openclaw_pipeline`:

- `run_id`
- `output_dir`
- `raw_output`
- `delivery_output`
- `report_output`
- `digest_brief_output`
- `pulled_count`
- `delivered_count`
- `marked_read_count`
- `status_counts`
- `delivery_payload`
- `keyword_index`

Optional:

- `items`
  - returned only when `include_item_reports=true`

Follow-up structured reads should use MCP tools rather than re-reading these files directly.

## Minimal Output Files

By default the pipeline writes these core artifacts:

- `outputs/freshrss/rerun/<run_dir>/run-state.json`
- `outputs/freshrss/rerun/<run_dir>/raw/freshrss.raw.json`
- `outputs/freshrss/rerun/<run_dir>/candidates/openclaw-delivery-payload.json`
- `outputs/freshrss/rerun/<run_dir>/candidates/digest-brief.json`
- `outputs/freshrss/rerun/<run_dir>/run-report.json`
- `outputs/freshrss/rerun/<run_dir>/extracted/item-XX.extracted.json` (one per item)

It also updates local runtime keyword data:

- `data/term_index/daily/YYYY-MM-DD.json`
- `data/term_index/term_stats.json`

Per-item extracted files live under `extracted/` and are always written.
If `debug_artifacts=true`, the pipeline additionally writes normalized items, summaries, filter decisions, candidate records, and candidate inputs.
The main pipeline does not emit a batch-level `freshrss.extracted.json` file by default.

## `resume_run` Minimal Scope

`resume_run` currently supports only the minimum resume contract:

- only runs with a valid `run-state.json`
- only workflow `freshrss_daily_digest`
- resume in place on the original `run_id`
- supported resume points: `generate_summaries`, `apply_filters`, `build_delivery_payload`, `write_run_report`
- unsupported resume points: `fetch_feed`, `extract_articles`
- if required artifacts are missing, the tool returns a non-resumable response instead of silently falling back to an earlier stage

Artifact expectations by resume point:

- `generate_summaries`: requires `raw/freshrss.raw.json` and `extracted/`
- `apply_filters`: requires the above plus per-item summary outputs
- `build_delivery_payload`: requires per-item candidate inputs consistent with filter-stage output
- `write_run_report`: requires `candidates/openclaw-delivery-payload.json`; if `mark_read=true`, raw input must still be present

## Payload Specs

OpenClaw payload field specs live here:

- `docs/openclaw/openclaw-candidate-input-field-spec.md`
- `docs/openclaw/openclaw-delivery-payload-spec.md`

## Read-State Semantics

The pipeline reads from FreshRSS unread items by default.

If `mark_read=true`:

- items are marked as read only after the final `openclaw-delivery-payload.json` has been written successfully
- only successfully delivered items are marked as read
- failed or skipped items remain unread

## FreshRSS Content Policy

For FreshRSS items, the pipeline is RSS-first and does not re-crawl webpages.

Behavior:

- use `item.raw_content` first
- if missing, use `item.raw_summary`
- if neither contains usable content, skip the item
- do not fetch the original webpage again for FreshRSS items

This is intentional.

## Keyword Cleanup Governance

This repository also includes a lightweight keyword-governance flow for downstream review.

Current pieces:

- runtime keyword stats
  - `data/term_index/daily/YYYY-MM-DD.json`
  - `data/term_index/term_stats.json`
- governance config
  - `configs/term_cleanup_policy.json`
  - `configs/term_watchlist.json`
  - `configs/term_change_log.json`
- review bundle builder
  - `skills/keyword-cleanup-review/scripts/build_review_bundle.py`
- accepted-suggestion writer
  - `scripts/apply_term_suggestions.py`

Current status:

- OpenClaw can read the keyword review bundle as maintenance input
- accepted suggestions still require explicit human confirmation
- the repository can write accepted watch / alias / stopword / interest-keyword changes after confirmation
- this governance flow is not yet wired into a periodic scheduler inside the repository

Boundary:

- keyword cleanup is a maintenance flow, not the production RSS ingestion path
- the repository does not auto-apply cleanup suggestions without confirmation
- current keyword stats are built from the delivered candidate payload, not yet from a final `DailyDigest`

## Known Limitations

- Some sources put only partial content in RSS; those items may be skipped if RSS content is insufficient.
- WeChat articles often block direct crawling, but this pipeline now avoids that path for FreshRSS items and uses RSS-provided content when available.
- Rule behavior is still conservative in some cases; many items may land in `review` depending on current rules.
- Paywall heuristics may produce false positives for some Chinese text patterns.
- Keyword cleanup governance is usable now, but periodic scheduling and before/after evaluation are not implemented yet.

## Files OpenClaw Should Read First

Recommended reading order for a new maintainer:

1. `README.md`
2. `docs/openclaw/openclaw-handoff.md`
3. `docs/openclaw/openclaw-candidate-input-field-spec.md`
4. `docs/openclaw/openclaw-delivery-payload-spec.md`
5. `docs/design/daily-keyword-index-design.md`
6. `skills/keyword-cleanup-review/SKILL.md`
7. `docs/current/context-reset-brief.md`

## Downstream Boundary Rules

For the daily-digest workflow:

- the digest should go to Hugo and chat reporting, not directly into IMA
- the full daily digest should **not** be uploaded to IMA
- only explicitly user-selected article summaries should be uploaded to IMA
- selected-article summaries should be generated from existing extracted text, not by re-fetching original URLs

## Current Recommendation

For integration handoff, the repository is usable now.

The minimum you need to give OpenClaw is:

- the repository code
- the MCP server startup command
- the required environment variables in the target environment
- the instruction to call `run_freshrss_openclaw_pipeline`
- the rule that follow-up state/result reads must go through MCP tools first, not handwritten filesystem paths

If OpenClaw will also participate in keyword-governance review, additionally point it to:

- `docs/design/daily-keyword-index-design.md`
- `skills/keyword-cleanup-review/SKILL.md`
- `scripts/apply_term_suggestions.py`