Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add epicsagas/plugins --skill reflectgit clone --depth 1 https://github.com/epicsagas/pluginsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/epicsagas/plugins/reflect)<a href="https://agentmods.dev/skills/epicsagas/plugins/reflect"><img src="https://agentmods.dev/badge/skills/epicsagas/plugins/reflect/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/epicsagas/plugins/reflect"><img src="https://agentmods.dev/badge/skills/epicsagas/plugins/reflect.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00067 | $0.00541 |
| Opus 5.5 | $0.00027 | $0.00216 |
| Sonnet 5.5 | $0.00013 | $0.00108 |
| Haiku 4.5 | $0.00007 | $0.00054 |
Grade A, and why
reflect scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
Self-assessment of AI session quality across 5 dimensions. Uses harness session data and memory to produce an honest scorecard.
Usage
/reflect
The 5 dimensions
| Dimension | Weight | Measures |
|---|---|---|
| Goal clarity | 20% | Were requirements specific enough? Did the user provide clear acceptance criteria? |
| Prompt quality | 20% | Were instructions precise? Did the user catch misunderstandings early? |
| Iteration efficiency | 20% | How many correction cycles? Were they due to unclear specs or exploration? |
| Tool utilization | 20% | Did the user leverage available tools (memory, search, code analysis)? |
| Outcome alignment | 20% | Does the final result match what was actually needed? |
Process
- Collect data —
harness_reflect_contextto get session stats and evolution history - Recall context —
harness_mem_recallwith hint about the current session - Score each dimension — 1-5 scale with specific evidence
- Calculate overall — weighted average
- Identify top improvement — single most impactful thing to do differently
- Record —
harness_mem_addthe reflection as type "insight"
Output format
## Session Reflection
### Scores
| Dimension | Score | Evidence |
|---|---|---|
| Goal clarity | 4/5 | ... |
| Prompt quality | 3/5 | ... |
| Iteration efficiency | 4/5 | ... |
| Tool utilization | 2/5 | ... |
| Outcome alignment | 5/5 | ... |
| **Overall** | **3.6/5** | |
### What went well
- ...
### What to improve
- ...
### Top recommendation for next session
> Single most impactful change to make.
Important notes
- This is NOT about agent performance — it's about how the human used AI
- Be honest, not flattering — low scores are more useful than high scores
- The "top recommendation" should be actionable and specific
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago First seen · 67 lines · 67 tokens per session scan A 82fef61c8473
reflect is a skill published in the GitHub repository epicsagas/plugins (2 stars, last pushed 18d ago), licensed Apache-2.0. It adds 67 tokens to every session and 541 once invoked, about $0.0003 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-05.
Other skills, from other repositories
hold-my-hand
Standing posture for multi-step work: only the current phase (exact steps, expected output, common failure, who acts, advancing reply), later ones as a 'Phase 1 of 8' count. Use when: 'hold my hand', 'walk me through this', 'one step at a time', 'one phase at a time', 'don't show me the next step yet', 'guide me…
prerequisites
Read-only report of external tools the enabled plugin fleet declares, and which of them are missing. Use when a hook says a formatter or CLI is missing, or when asking whether this machine has what the enabled plugins need. Never installs.
progress
A status-checking workflow for finding out whether programs or jobs are still running, have finished, or stopped unexpectedly.
systematic-debugging
Systematic debugging approach for errors, test failures, and unexpected behavior. Use when encountering bugs, test failures, errors, or when debugging is needed. Prevents common anti-patterns like random fixes, skipping root cause analysis, and thrashing.
product-pulse
Use to produce a single-page dated pulse report on product health — usage metrics, AI-feature signals (hallucination, refusal, eval drift), counter-metrics, follow-ups. Trigger phrases include "run a pulse on [area]", "give me a product health snapshot", "weekly product pulse", "what does the data say this week".…
patterns-watch
Use to surface unsolicited patterns across stakeholder reflections, wins log, and self-reflect entries. Trigger phrases include "what patterns am I missing", "anything notable across my files", "scan my reflections for things I should know", "weekly cross-cutting check". Distinct from stakeholder-synthesize…