sawz/last30days-skill
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat(release): v2.9.4 - move save into Python script, zero post-invitation noise

Browse Source 
Add --save-dir flag to last30days.py that saves raw research output
during the existing script run. Remove entire "Save Research to
Documents" section from SKILL.md (~45 lines). No more extra tool
calls, no (No output), no multi-minute cogitation after invitation.

Tested: --mock confirms file creation and duplicate date suffixing.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This commit is contained in:
Matt Van Horn
2026-03-06 16:27:53 -08:00
parent 28d223e43c
commit 6d5acb9121
4 changed files with 116 additions and 53 deletions
Download Patch File Download Diff File Expand all files Collapse all files
CHANGELOG.md
+8
View File
@@ -5,6 +5,14 @@ All notable changes to this project will be documented in this file.
The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
## [2.9.4] - 2026-03-06
### Changed
- Move save into Python script via `--save-dir` flag - raw research data saved during the existing script Bash call, zero extra tool calls after invitation
- Remove entire "Save Research to Documents" section from SKILL.md (~45 lines removed)
- No more `📎` footer, no Bash heredoc, no `(No output)`, no multi-minute cogitation after research
## [2.9.3] - 2026-03-06
### Fixed
SKILL.md
+4 -53
View File
@@ -30,7 +30,7 @@ metadata:
- prompts
---
# last30days v2.9.3: Research Any Topic from the Last 30 Days
# last30days v2.9.4: Research Any Topic from the Last 30 Days
Research ANY topic across Reddit, X, YouTube, TikTok, Hacker News, Polymarket, and the web. Surface what people are actually discussing, recommending, betting on, and debating right now.
@@ -123,7 +123,7 @@ If `--agent` appears in ARGUMENTS (e.g., `/last30days plaud granola --agent`):
5. **Skip** the follow-up invitation ("I'm now an expert on X...")
6. **Output** the complete research report and stop - do not wait for further input
Agent mode still saves the research briefing to `~/Documents/Last30Days/` using the same logic as interactive mode (see "Save Research to Documents" section).
Agent mode saves raw research data to `~/Documents/Last30Days/` automatically via `--save-dir` (handled by the script, no extra tool calls).
Agent mode report format:
@@ -167,7 +167,7 @@ if [ -z "${SKILL_ROOT:-}" ]; then
exit 1
fi
python3 "${SKILL_ROOT}/scripts/last30days.py" "$ARGUMENTS" --emit=compact --no-native-web # Add --x-handle=HANDLE if RESOLVED_HANDLE is set
python3 "${SKILL_ROOT}/scripts/last30days.py" "$ARGUMENTS" --emit=compact --no-native-web --save-dir=~/Documents/Last30Days # Add --x-handle=HANDLE if RESOLVED_HANDLE is set
```
Use a **timeout of 300000** (5 minutes) on the Bash call. The script typically takes 1-3 minutes.
@@ -498,58 +498,9 @@ For `/last30days war in Iran` (NEWS):
---
## Save Research to Documents
After displaying the invitation, save the complete research briefing to `~/Documents/Last30Days/`. This happens automatically on every run.
**Generate the filename** from TOPIC:
- Lowercase, replace spaces/special chars with hyphens, remove consecutive hyphens, trim to 60 chars
- Example: "Claude Code Best Practices" → `claude-code-best-practices.md`
- If file already exists, append today's date: `{slug}-YYYY-MM-DD.md`
**End your invitation with a single `📎` footer line:**
```
📎 ~/Documents/Last30Days/{slug}.md
```
**Then save using a foreground Bash command** (cat is instant - no delay):
```bash
mkdir -p ~/Documents/Last30Days && cat > ~/Documents/Last30Days/{slug}.md << 'RESEARCH_EOF'
# {TOPIC}
> Researched {date} | Query type: {QUERY_TYPE} | Target tool: {TARGET_TOOL or "general"}
## What I learned
{The full synthesis section you just displayed - all topics, patterns, and citations}
## Stats
{The full stats box with source counts and engagement - copy exactly as displayed}
## Follow-up suggestions
{The 2-3 specific suggestions from the invitation block}
---
*Generated by [last30days](https://github.com/mvanhorn/last30days-skill) v2.9.3*
RESEARCH_EOF
```
**CRITICAL RULES:**
1. NEVER use the `Write` tool - it displays "Wrote N lines..." which clutters the output
2. NEVER use `run_in_background: true` - background callbacks cause the model to re-engage, hallucinate fake user messages, and generate unsolicited responses
3. Use FOREGROUND Bash with `cat >` - it executes instantly (sub-second) and produces no output
4. The `📎` line is part of your text message, not a separate tool call
5. Do NOT output any text after the save Bash call. No "saved", no "ready for questions", nothing.
---
## WAIT FOR USER'S RESPONSE
**STOP and wait** for the user to respond. Do NOT generate any text after the save.
**STOP and wait** for the user to respond. Do NOT call any tools after displaying the invitation. The research script already saved raw data to `~/Documents/Last30Days/` via `--save-dir`.
---
docs/plans/2026-03-06-fix-save-output-noise-plan.md
+83
View File
@@ -0,0 +1,83 @@
---
title: "fix: Eliminate save-related output noise after research"
type: fix
status: active
date: 2026-03-06
---
# fix: Eliminate save-related output noise after research
## Problem Statement
After research completes, the auto-save feature adds 3-5 lines of unwanted output below the clean invitation block. Every approach tried so far has made things worse:
| Version | Approach | Lines added | Side effects |
|---------|----------|-------------|--------------|
| v2.9.1 | `Write` tool | ~5 | Shows "Wrote 43 lines to..." |
| v2.9.2 | `run_in_background: true` | ~5 | Background callback triggers 2-4 min cogitation, model hallucinates fake "Human:" messages and generates unsolicited multi-paragraph responses |
| v2.9.3 | Foreground `cat >` heredoc | ~3 | Shows `(No output)`, still triggers 2-4 min cogitation |
**Current v2.9.3 noise (3 lines):**
```
⏺ Bash(mkdir -p ~/Documents/Last30Days && cat > ...)
⎿ (No output)
✻ Churned for 2m 38s
```
The root cause is that ANY tool call after the invitation creates unavoidable Claude Code UI chrome, and the model cogitates for minutes regardless of foreground vs background.
## Proposed Solutions
### Option A: Remove auto-save entirely (Recommended)
Remove the entire "Save Research to Documents" section from SKILL.md. The research lives in the conversation. Zero extra tool calls, zero extra lines, zero cogitation.
**Changes to `SKILL.md`:**
- Delete the "Save Research to Documents" section (~45 lines)
- Update agent mode reference (line 126) - remove save mention
- Update security section (line 612) - change "Saves research briefings" to past tense or conditional
- Remove `📎` footer line from invitation format
- Simplify "WAIT FOR USER'S RESPONSE" section
**What users lose:** Auto-saved .md files in `~/Documents/Last30Days/`
**What users gain:** Clean output ending exactly at the invitation block
### Option B: Move save into Python script
Add `--save-dir` flag to `last30days.py`. The script saves its raw output during its existing Bash call (which already runs). Zero extra tool calls.
**Changes:**
- `scripts/last30days.py` - add `--save-dir` argument, write raw output to file at end of execution
- `SKILL.md` - add `--save-dir=~/Documents/Last30Days` to the script invocation, remove save section
**What users lose:** Saved file contains raw research data, not Claude's synthesis
**What users gain:** Zero extra lines, data is still preserved
### Option C: Opt-in save via follow-up command
Remove auto-save. Add a note to the invitation: "Say 'save' to save this research." When user says "save", run the Bash heredoc then.
**Changes to `SKILL.md`:**
- Delete auto-save section
- Add "save" as a recognized intent in "WHEN USER RESPONDS" section
- Save logic only runs when explicitly requested
**What users lose:** Nothing - save is still available on demand
**What users gain:** Clean output by default, save when they want it
## Acceptance Criteria
- [ ] Research output ends at the invitation block with zero tool calls after it
- [ ] No `(No output)` line visible after invitation
- [ ] No multi-minute cogitation after invitation
- [ ] No hallucinated fake user messages
- [ ] Version bumped to v2.9.4
- [ ] CHANGELOG updated
- [ ] Synced to ~/.claude, ~/.agents, ~/.codex via sync.sh
## Context
- Source: `/Users/mvanhorn/last30days-skill-private/SKILL.md`
- Deployed to: `~/.claude/skills/last30days/SKILL.md`
- Sync command: `bash scripts/sync.sh`
- Current version: v2.9.3
scripts/last30days.py
+21
View File
@@ -1250,6 +1250,13 @@ def main():
default=False,
help="Skip native web search backends (Parallel/Brave/OpenRouter). Use when the assistant has its own WebSearch tool.",
)
parser.add_argument(
"--save-dir",
type=str,
default=None,
metavar="DIR",
help="Auto-save raw research output to DIR/{topic-slug}.md",
)
args = parser.parse_args()
args.topic = " ".join(args.topic) if args.topic else None
@@ -1604,6 +1611,20 @@ def main():
# Output result
output_result(report, args.emit, web_needed, args.topic, from_date, to_date, missing_keys, args.days, source_info)
# Auto-save raw research to file if --save-dir is set
if args.save_dir:
import re
save_dir = Path(args.save_dir).expanduser()
save_dir.mkdir(parents=True, exist_ok=True)
slug = re.sub(r'[^a-z0-9]+', '-', args.topic.lower()).strip('-')[:60]
save_path = save_dir / f"{slug}.md"
if save_path.exists():
save_path = save_dir / f"{slug}-{datetime.now().strftime('%Y-%m-%d')}.md"
content = render.render_compact(report, missing_keys=missing_keys)
content += "\n" + render.render_source_status(report, source_info)
save_path.write_text(content, encoding="utf-8")
print(f"📎 {save_path}", file=sys.stderr)
# Persist findings to SQLite if requested
if args.store:
import store as store_mod