reground

reground is a skill for Claude Code from adamlinscott/claude-skills. It costs 95 tokens per session (981 once invoked), scanned A, original, MIT.

A skill that stops an agent that has started drifting away from the evidence in the codebase and the task. It clears speculative assumptions before work continues.

In plain words
What is it for?
Use it to re-anchor an agent to the actual repository and current task after it goes off course.
Why use it?
It helps when an agent invents files or APIs, follows an incorrect idea, or starts building more than requested.

Skill for Claude Code

Written for Claude Code: shipped in a Claude Code plugin.

Part of the claude-skills plugin — 18 skills shipped together

Good fit Use it to re-anchor an agent to the actual repository and current task after it goes off course.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/adamlinscott/claude-skills/reground
Install

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.

Any agent
npx skills add adamlinscott/claude-skills --skill reground
Clone the repo
git clone --depth 1 https://github.com/adamlinscott/claude-skills

Made for: Claude Code.

Or install claude-skills, the plugin that ships this one along with the rest of its 18 skills.

Wrote 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.

agentmods badge for reground

README.md
[![agentmods](https://agentmods.dev/badge/skills/adamlinscott/claude-skills/reground/github.svg)](https://agentmods.dev/skills/adamlinscott/claude-skills/reground)
Your own site
<a href="https://agentmods.dev/skills/adamlinscott/claude-skills/reground"><img src="https://agentmods.dev/badge/skills/adamlinscott/claude-skills/reground/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.

agentmods 80×15 button for reground

Your own site · 80×15
<a href="https://agentmods.dev/skills/adamlinscott/claude-skills/reground"><img src="https://agentmods.dev/badge/skills/adamlinscott/claude-skills/reground.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 95 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 981 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00095 $0.00981
Opus 5 $0.00048 $0.00491
Sonnet 5 $0.00019 $0.00196
Haiku 4.5 $0.00010 $0.00098

Measured 9d ago against content hash d4b179d7add0, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-09, from the pricing page.

Security

Grade A, and why

reground 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 9d 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.

skills/reground/SKILL.md · 74 lines

How it starts

The opening of the file, as written. The whole thing — 74 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Reground

The current session has drifted: the recent work rests on assumptions the codebase has not been shown to support, invents systems that may not exist, or builds past what the task needs. This skill is a hard reset to ground truth. It does not summarize or compact — it discards the speculative thread and rebuilds a small, evidence-backed footing for the next step. Run the protocol in order. Do not skip ahead to step 5.

1. Halt

Stop immediately. Make no further Edit, Write, or other tool calls that advance the work until the report in step 5 is produced and the user approves a next step.

Before reporting, make sure the working tree is in a coherent state — never leave a file half-written or a multi-file change partly applied. If a change is mid-flight, the only edits permitted here are the minimal ones that restore a self-consistent, non-broken state, or a clean revert of it; make no new feature progress either way. The aim is solid footing to assess from, not a frozen half-edit.

2. Name the drift

State plainly, in one or two sentences, the specific unverified assumption(s) the recent work depended on — the equivalent of "the last answer assumed X, but the repo shows no evidence for X." Be concrete: name the assumed file, function, endpoint, schema, or behavior. If you cannot name what was assumed, that itself is the finding — say so.

3. Set the anchor

Decide the focus file set — the minimum slice of the repo the next step actually depends on:

  • If the user passed paths or class/function/object names (as arguments or in their message), use exactly those.
  • Otherwise, self-derive them from the stated task: list the files/paths you intend to read and why each one matters, then read them. Keep the set small and task-scoped; do not sweep the whole repo. If the right anchor is genuinely unclear, ask the user for the paths rather than guessing.

4. Read the anchor, nothing else

Read those files now. Treat only what is on disk as true. Do not fill gaps with memory, convention, or inference about code you have not opened. If something the task seems to need is absent, that is an "unknown," not a thing to assume into existence.

Read the full file on GitHub · 74 lines

Changes

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.

  1. 9d ago First seen · 74 lines · 95 tokens per session scan A d4b179d7add0

Subscribe to this mod's changes

reground is a skill published in the GitHub repository adamlinscott/claude-skills (3 stars, last pushed 9d ago), licensed MIT. It adds 95 tokens to every session and 981 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.