daily-reflection
Scheduled end-of-day synthesis. Runs at 18:00 daily — reads the day's sessions, writes a structured journal entry to memory/journal/, and proposes rule candidates from patterns across recent journals. Use when the harness schedule fires or when the user asks to summarize today's activity.
- License 许可
- MIT
- 最近更新
- 2026-04-30
- Stars
- ★ 1
- Forks
- 2
01 前言 Front Matter
- name
- daily-reflection
- license
- MIT
- metadata
- { "harness-tags": "workflow,scheduled,reflection", "harness-author": "human", "harness-status": "active", "harness-related": "research,respect-the-user", "harness-schedule": "0 18 * * *" }
- description
- Scheduled end-of-day synthesis. Runs at 18:00 daily — reads the day's sessions, writes a structured journal entry to memory/journal/, and proposes rule candidates from patterns across recent journals. Use when the harness schedule fires or when the user asks to summarize today's activity.
02 资源文件 8
- 参考文档evals/evals.json1.4 KB
- 参考文档evals/files/correction-pattern.md943 B
- 参考文档evals/files/session-1.md668 B
- 参考文档evals/files/session-2.md677 B
- 参考文档evals/files/uneventful.md243 B
- 参考文档evals/triggers.json3.0 KB
- 脚本scripts/propose-rules.sh5.3 KB
- 脚本scripts/synthesize.sh6.7 KB
03 SKILL.md 正文
Workflow: Daily Reflection
When to use
This skill is triggered automatically at 18:00 by the harness scheduler
(metadata.harness-schedule: 0 18 * * *). Invoke it manually when:
- The user asks to synthesize or summarize today's work.
- The user asks what was accomplished, learned, or left open.
- You want to extract rule candidates from recent journals after a significant session.
Available scripts
scripts/synthesize.sh [--date YYYY-MM-DD] [--harness-dir <path>]— Run the daily journal synthesis for the given date (default: today). Returns structured JSON withjournal_path,sessions_processed, andpatterns_detected.scripts/propose-rules.sh [--harness-dir <path>]— Analyze recent journals and propose instinct/rule candidates. Returns JSON with a list of candidates and their source reasoning.
Workflow
- Synthesize: Run
scripts/synthesize.sh(optionally with--date). Onstatus: ok, the journal file is atresult.journal_path. - Propose rules: Run
scripts/propose-rules.shto extract rule candidates from the new (and recent) journal entries. - Review: Present the journal summary and rule candidates to the user for review before any rule is promoted to an instinct.
Gotchas
- Both scripts require a working
harnessbinary onPATH(or--harness-dirpointing to a valid harness directory). synthesize.shis idempotent by default — re-running on the same date overwrites the existing journal entry. Pass--no-forceif you want to preserve an existing entry.propose-rules.shreads the last 7 days of journal entries. On a fresh harness with fewer than 7 days of history the output will be sparse — this is expected.
Failure modes
NO_SESSIONS— No session files found for the requested date. Nothing to synthesize. Inform the user and exit.JOURNAL_COMMAND_FAILED— The underlyingharness journalcommand returned a non-zero exit code. The error details are inerror.message. Check that Ollama is running and the configured model is available.LEARN_COMMAND_FAILED— Theharness learncommand failed. Seeerror.message. Same checks as above.INVALID_DATE— The--dateargument does not matchYYYY-MM-DD. Fix the format and retry.
04 同仓库更多 Skills
research
Conduct research.
{{NAME=null}=null}
Briefly describe what this skill does AND when to use it. Use imperative phrasing ("Use this skill when..."). Aim for 80-500 chars. Include keywords the model can pattern-match against the user's prompt.
working-in-a-git-repo
Operating manual for dev work: worktree discipline, build/test/lint commands, file ownership zones, commit style, and a verification checklist before reporting done. Use when starting any feature/fix work in this repo.
test-runner
Runs the project's test suite (or a targeted subset), parses the output, and reports pass/fail with failing-test details. Use when you need to know test status; call with the test command to run, or leave blank for the project default.
reviewer
Reviews a diff or change set for bugs, security issues, test gaps, and style. Use when evaluating a code change, reviewing a pull request, or running a structured review pass before merge.
fix-failing-tests
Seven-step flow for diagnosing and fixing failing tests. Use when the user asks to fix failing tests, green the suite, or says 'tests are red' — works through reproduce → read → root-cause → propose → apply → verify → commit.