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 agentmods add skills/stealthy-mcstealth/self-evolve/reload-contextnpx skills add Stealthy-McStealth/self-evolve --skill reload-contextgit clone --depth 1 https://github.com/Stealthy-McStealth/self-evolveWrote 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/stealthy-mcstealth/self-evolve/reload-context)<a href="https://agentmods.dev/skills/stealthy-mcstealth/self-evolve/reload-context"><img src="https://agentmods.dev/badge/skills/stealthy-mcstealth/self-evolve/reload-context.svg" alt="Measured on agentmods" 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 | $0.00102 | $0.00530 |
| Opus 5 | $0.00051 | $0.00265 |
| Sonnet 5 | $0.00020 | $0.00106 |
| Haiku 4.5 | $0.00010 | $0.00053 |
Grade A, and why
reload-context 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 5d 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
reload-context
Long conversations cause harness drift. You loaded a skill twenty turns ago and now you're winging it. This skill fixes that.
The research is clear: adherence to loaded instructions decays over trajectories. Strong models drift less; all models drift some. This is the manual override.
Step 1: Identify what's loaded
List every skill, instruction, or rule that is currently active in this session:
- Skills explicitly loaded or invoked
- Any CLAUDE.md instructions in scope
- Project-specific conventions established earlier in this conversation
- Any explicit rules the user stated ("always do X", "never do Y")
Step 2: Re-read each one
Don't summarize from memory. Actually re-read each skill file or instruction set. If a skill file path is known, use Read to load the current content — it may have been updated since you first saw it.
Step 3: State what you're now following
Write a short, explicit list of the active constraints and procedures:
reloaded. currently following:
- [skill-name]: [the key procedure in one line]
- [rule from user]: [the rule]
- [project convention]: [the convention]
This makes the reload visible and auditable. The user can correct anything that's wrong or missing.
Step 4: Commit to the fallbacks
For any procedural skill that has fallback steps: explicitly acknowledge them. "If step A fails, I will try step B before giving up" — not just "I will follow the skill."
The most common failure mode is treating step 1 as the whole procedure. Reloading means recommitting to the full procedure including fallbacks.
What this doesn't do
This skill re-anchors to existing loaded instructions. It doesn't:
- Create new skills (use
/improve-urselffor that) - Erase conversation history
- Change your base behavior
It's a deliberate attention reset. That's all it needs to be.
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 5d ago First seen · 74 lines · 102 tokens per session scan A 1f0719ca8bbf
reload-context is a skill published in the GitHub repository Stealthy-McStealth/self-evolve (5 stars, last pushed 3mo ago), licensed MIT. It adds 102 tokens to every session and 530 once invoked, about $0.0005 per session on Opus 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 2026-08-31.
Other skills, from other repositories
debug-memory
Reproduce observational-memory bugs (reflection, recall, observer, freshness, eviction) by dumping the live /.duet/memory.db into a fixture and driving an eval against it. Use whenever the agent's memory misbehaves — wrong observations, bad reflections, missing recall, runaway tokens — or whenever the user asks to…
write-eval
Write a live eval for new or changed runner/agent behavior using red/green TDD plus a falsification check that proves the eval fails when the behavior is broken. Use whenever you add or modify behavior that should be covered by an eval, when asked to "write an eval", "add an eval", "cover this with an eval", or after…
compare-screenshots
Compare screenshots to judge which image is less wrong, not to match a baseline. Use when a UI, game, document, render, chart, or generated asset needs objective visual telemetry, side-by-side inspection, crop/zoom review, or a fresh second opinion before accepting or rejecting a visual change.
release
Increment the package version, commit the release, create a matching version tag, and push main plus the tag. Use when releasing this repo, publishing the package, bumping the version, tagging a release, or when the user asks to "release", "bump ver", "tag", or "push tags".
refactor-clean
Refactor cleanly instead of layering sediment. Use when a change reveals duplicated concepts, local adapters, obsolete owners, compatibility wrappers, parallel abstractions, an over-large module that has accreted many responsibilities, or "just tack this on" pressure in any code area.
screenshot-critique
Use the unprimed sub agent as a second set of eyes before accepting visual work — MANDATORY before declaring any user-reported visual bug fixed or claiming a visual change verified; primed eyes pass defects fresh eyes catch.