How do I install the Improve Context skill?

Run npx skillmds@latest add ai-analyst-lab/improve-context in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Improve Context safe to use?

This skill has not completed SkillMD's automated safety review yet. Capability flags: docs only. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Improve Context?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Improve Context free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Improve Context?

ai-analyst-lab (@ai-analyst-lab) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Improve Context by ai-analyst-lab — Agent Skill

Name: Improve Context
Rating: 4.2 (5 reviews)
Author: ai-analyst-lab

Improve Context

Diagnose a context-related analytical failure, propose the smallest relevant change, compare the same evaluation cases before and after, and produce an accept, revise, or revert receipt. Use when the user asks to improve context, fix a definition or example, test whether context helped, or invokes /improve-context.

File contents

Improve context

Use the evaluation system as the referee. Do not promote a context change because the new answer looks more plausible.

1. Diagnose

Read the failing evaluation record and its analysis trace. Classify the context failure:

missing information;
wrong information;
stale information;
conflicting information;
correct information was not supplied;
supplied information was ignored or misapplied;
the problem belongs in a skill, helper, workflow, or evaluator rather than context.

Run /context-trace on the exact question. Preserve its JSON fingerprint.

2. Propose

Create working/context-change-proposal.yaml with:

observed failure and evidence;
smallest proposed change;
representation and delivery;
source and owner;
status proposed;
cases expected to improve;
possible regressions;
required reviewers;
acceptance criteria.

Do not change trusted context until the user approves the proposal.

3. Preserve the baseline

Run the working and regression cases before the edit. Keep the manifest paths and run IDs. Freeze the suite, data snapshot, model, evaluator, and trial count. Record the proposal's named target case IDs. For the Session 8 course fallback, use data/evals/public/week4-context.yaml with the context-policy runner. Lock outputs before grading against data/evals/working-references/week4-context.yaml.

4. Apply one change

Make only the approved context change. Show the diff. Do not change the evaluator or expected results in the same comparison.

5. Rerun and inspect

Run the same working and regression cases. Inspect:

aggregate score;
target cases;
regressions;
cases still failing;
context manifest;
context citations and SQL-use evidence.

6. Write the receipt

Use helpers.knowledge.context_change with the baseline and candidate manifests. Pass every named target with --target-case, and verify that the actual changed system paths are a subset of the approved changed paths. The receipt must recommend:

accept when target cases improve, no passing case regresses, and the comparison is controlled;
revert when a passing case regresses without an understood exception;
revise when the runs are not comparable or the change does not improve the target behavior.

The user chooses whether to promote the proposal. The system may draft the change and evidence, but it does not ratify business truth.

---name: improve-contextdescription: Diagnose a context-related analytical failure, propose the smallest relevant change, compare the same evaluation cases before and after, and produce an accept, revise, or revert receipt. Use when the user asks to improve context, fix a definition or example, test whether context helped, or invokes /improve-context.---# Improve contextUse the evaluation system as the referee. Do not promote a context change because the new answer looks more plausible.## 1. DiagnoseRead the failing evaluation record and its analysis trace. Classify the context failure:- missing information;- wrong information;- stale information;- conflicting information;- correct information was not supplied;- supplied information was ignored or misapplied;- the problem belongs in a skill, helper, workflow, or evaluator rather than context.Run `/context-trace` on the exact question. Preserve its JSON fingerprint.## 2. ProposeCreate `working/context-change-proposal.yaml` with:- observed failure and evidence;- smallest proposed change;- representation and delivery;- source and owner;- status `proposed`;- cases expected to improve;- possible regressions;- required reviewers;- acceptance criteria.Do not change trusted context until the user approves the proposal.## 3. Preserve the baselineRun the working and regression cases before the edit. Keep the manifest paths and run IDs. Freeze the suite, data snapshot, model, evaluator, and trial count.Record the proposal's named target case IDs. For the Session 8 course fallback, use`data/evals/public/week4-context.yaml` with the `context-policy` runner. Lock outputs before gradingagainst `data/evals/working-references/week4-context.yaml`.## 4. Apply one changeMake only the approved context change. Show the diff. Do not change the evaluator or expected results in the same comparison.## 5. Rerun and inspectRun the same working and regression cases. Inspect:- aggregate score;- target cases;- regressions;- cases still failing;- context manifest;- context citations and SQL-use evidence.## 6. Write the receiptUse `helpers.knowledge.context_change` with the baseline and candidate manifests. Pass every namedtarget with `--target-case`, and verify that the actual changed system paths are a subset of theapproved changed paths. The receipt must recommend:- accept when target cases improve, no passing case regresses, and the comparison is controlled;- revert when a passing case regresses without an understood exception;- revise when the runs are not comparable or the change does not improve the target behavior.The user chooses whether to promote the proposal. The system may draft the change and evidence, but it does not ratify business truth.

Frequently asked questions