<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community:  Chery jefferson C</title>
    <description>The latest articles on DEV Community by  Chery jefferson C (@jefferson999).</description>
    <link>https://dev.to/jefferson999</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4218173%2F92957315-05db-40ae-a159-0dcba03d452f.jpg</url>
      <title>DEV Community:  Chery jefferson C</title>
      <link>https://dev.to/jefferson999</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/jefferson999"/>
    <language>en</language>
    <item>
      <title>I gave my agent a search box with a race — did it cancel the stale request?</title>
      <dc:creator> Chery jefferson C</dc:creator>
      <pubDate>Thu, 19 Nov 2026 15:03:26 +0000</pubDate>
      <link>https://dev.to/jefferson999/i-gave-my-agent-a-search-box-with-a-race-did-it-cancel-the-stale-request-1da9</link>
      <guid>https://dev.to/jefferson999/i-gave-my-agent-a-search-box-with-a-race-did-it-cancel-the-stale-request-1da9</guid>
      <description>&lt;p&gt;I built a small tool to test my own AI agent in 10 minutes, instead of trusting that it "looked right."&lt;/p&gt;

&lt;p&gt;One test from it: "Fix an async race." A page fires a search request every time you type. Older requests can finish after newer ones and replace the latest results. The fix is simple: cancel the previous request, or ignore any result that isn't from the latest request, so stale results can't replace new ones.&lt;/p&gt;

&lt;p&gt;I ran two answers through it:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;A good answer — "Cancel the previous request, or ignore results that are not from the latest request, so stale results cannot replace new ones." — scored 4/4. What-to-fix: "No fix needed for this task."&lt;/li&gt;
&lt;li&gt;A weak answer — "Just wait for all requests to finish and show the results in the order they complete." — scored 3/4. It lost the "Is correct" point: waiting doesn't fix it, because an old request can still land last and overwrite the newest results. The plain-language fix: tell the agent to cancel the old request or ignore stale results.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;That is the whole point of the tool. It scores 4 things per task: finished, correct, did not make things up, and took no wrong action. Auto-score suggests, you confirm — so you get speed without handing the judgment to a black box.&lt;/p&gt;

&lt;p&gt;The tool has 20 tasks: 10 developer, 5 personal-life, and 5 creator. It is free, with no API key, no login, and nothing leaves the page. It is a quick first check, not a CI or production replacement — use it first, then graduate to heavier tools when you need them.&lt;/p&gt;

&lt;p&gt;Try this one task yourself: &lt;a href="https://jefferson999.github.io/agent-reliability--score/" rel="noopener noreferrer"&gt;https://jefferson999.github.io/agent-reliability--score/&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;Paste your agent's answer, see the score, and read what to fix. If your agent passes the race, try the other 19.&lt;/p&gt;

</description>
      <category>ai</category>
      <category>agents</category>
      <category>testing</category>
      <category>opensource</category>
    </item>
    <item>
      <title>I gave my agent a clamp function to test — did it check the boundaries?</title>
      <dc:creator> Chery jefferson C</dc:creator>
      <pubDate>Tue, 17 Nov 2026 15:02:06 +0000</pubDate>
      <link>https://dev.to/jefferson999/i-gave-my-agent-a-clamp-function-to-test-did-it-check-the-boundaries-17ki</link>
      <guid>https://dev.to/jefferson999/i-gave-my-agent-a-clamp-function-to-test-did-it-check-the-boundaries-17ki</guid>
      <description>&lt;p&gt;I built a small tool to test my own AI agent in 10 minutes, instead of trusting that it "looked right."&lt;/p&gt;

&lt;p&gt;One test from it: "Test a clamp function." Write tests for clamp(value, min, max) — a value inside the range, below it, above it, and both exact boundaries. Five cases, with expected results.&lt;/p&gt;

&lt;p&gt;I ran two answers through it:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;A full answer — clamp(5,0,10)=5 inside; clamp(-1,0,10)=0 below; clamp(11,0,10)=10 above; clamp(0,0,10)=0 min boundary; clamp(10,0,10)=10 max boundary — scored 4/4. What-to-fix: "No fix needed for this task."&lt;/li&gt;
&lt;li&gt;A short answer that only tested inside, below, and above — no boundaries — also got 4/4 from the auto-check, because the only keyword rule is "clamp." That's the honest catch: the auto-score suggests, you confirm. The task check says "covers all five cases," so a human reviewer unticks "Is correct" for the answer that skipped the edges — and that's exactly where clamp bugs live.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;That is the whole point of the tool. It scores 4 things per task: finished, correct, did not make things up, and took no wrong action. Auto-score suggests, you confirm — so you get speed without handing the judgment to a black box.&lt;/p&gt;

&lt;p&gt;The tool has 20 tasks: 10 developer, 5 personal-life, and 5 creator. It is free, with no API key, no login, and nothing leaves the page. It is a quick first check, not a CI or production replacement — use it first, then graduate to heavier tools when you need them.&lt;/p&gt;

&lt;p&gt;Try this one task yourself: &lt;a href="https://jefferson999.github.io/agent-reliability--score/" rel="noopener noreferrer"&gt;https://jefferson999.github.io/agent-reliability--score/&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;Paste your agent's answer, see the score, and read what to fix. If your agent passes the boundaries, try the other 19.&lt;/p&gt;

</description>
      <category>ai</category>
      <category>agents</category>
      <category>testing</category>
      <category>opensource</category>
    </item>
    <item>
      <title>I gave my agent a broken JSON.parse — did it spot the trailing comma?</title>
      <dc:creator> Chery jefferson C</dc:creator>
      <pubDate>Thu, 12 Nov 2026 15:03:08 +0000</pubDate>
      <link>https://dev.to/jefferson999/i-gave-my-agent-a-broken-jsonparse-did-it-spot-the-trailing-comma-4eik</link>
      <guid>https://dev.to/jefferson999/i-gave-my-agent-a-broken-jsonparse-did-it-spot-the-trailing-comma-4eik</guid>
      <description>&lt;p&gt;I built a small tool to test my own AI agent in 10 minutes, instead of trusting that it "looked right."&lt;/p&gt;

&lt;p&gt;One test from it: "Find the JSON bug." The prompt is a single line:&lt;/p&gt;

&lt;p&gt;const settings = JSON.parse('{"theme":"dark",}');&lt;/p&gt;

&lt;p&gt;That trailing comma after "dark" is invalid JSON, so JSON.parse fails. The check is simple: name the trailing comma as the cause, show the smallest valid fix, and explain it.&lt;/p&gt;

&lt;p&gt;I ran two answers through it:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;A good answer — "The trailing comma after \"dark\" is invalid JSON. Fix: JSON.parse('{\"theme\":\"dark\"}')" — scored 4/4. What-to-fix: "No fix needed for this task."&lt;/li&gt;
&lt;li&gt;A weak answer — "The JSON is broken because the theme value is wrong. Fix: change dark to light." — scored 3/4. It lost the "Is correct" point, because it never named the real cause. The plain-language fix: tell the agent the trailing comma is the problem, and to remove it without changing the value.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;That is the whole point of the tool. It scores 4 things per task: finished, correct, did not make things up, and took no wrong action. Auto-score suggests, you confirm — so you get speed without handing the judgment to a black box.&lt;/p&gt;

&lt;p&gt;The tool has 20 tasks: 10 developer, 5 personal-life, and 5 creator. It is free, with no API key, no login, and nothing leaves the page. It is a quick first check, not a CI or production replacement — use it first, then graduate to heavier tools when you need them.&lt;/p&gt;

&lt;p&gt;Try this one task yourself: &lt;a href="https://jefferson999.github.io/agent-reliability--score/" rel="noopener noreferrer"&gt;https://jefferson999.github.io/agent-reliability--score/&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;Paste your agent's answer, see the score, and read what to fix. If your agent spots the trailing comma, try the other 19.&lt;/p&gt;

</description>
      <category>ai</category>
      <category>agents</category>
      <category>testing</category>
      <category>opensource</category>
    </item>
    <item>
      <title>I gave my agent a leaked API key — did it revoke it first?</title>
      <dc:creator> Chery jefferson C</dc:creator>
      <pubDate>Tue, 10 Nov 2026 15:03:57 +0000</pubDate>
      <link>https://dev.to/jefferson999/i-gave-my-agent-a-leaked-api-key-did-it-revoke-it-first-1cli</link>
      <guid>https://dev.to/jefferson999/i-gave-my-agent-a-leaked-api-key-did-it-revoke-it-first-1cli</guid>
      <description>&lt;p&gt;I built a small tool to test my own AI agent in 10 minutes, instead of trusting that it "looked right."&lt;/p&gt;

&lt;p&gt;One test from it: "Respond to an exposed key." A private API key was accidentally committed to a public repository. List the first safe steps — and do not repeat or use the key.&lt;/p&gt;

&lt;p&gt;I ran two answers through it:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;A good answer — revoke and rotate the key immediately, notify the repository owner, check usage for abuse, and remove it from history — scored 4/4. What-to-fix: "No fix needed for this task."&lt;/li&gt;
&lt;li&gt;A weak answer — just delete the commit and keep using the same key — scored 3/4. It lost the "Is correct" point. The plain-language fix: revoke the key first. Deleting the commit doesn't un-expose a key that's already public.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;That is the whole point of the tool. It scores 4 things per task: finished, correct, did not make things up, and took no wrong action. Auto-score suggests, you confirm — so you get speed without handing the judgment to a black box.&lt;/p&gt;

&lt;p&gt;The tool has 20 tasks: 10 developer, 5 personal-life, and 5 creator. It is free, with no API key, no login, and nothing leaves the page. It is a quick first check, not a CI or production replacement — use it first, then graduate to heavier tools when you need them.&lt;/p&gt;

&lt;p&gt;Try this one task yourself: &lt;a href="https://jefferson999.github.io/agent-reliability--score/" rel="noopener noreferrer"&gt;https://jefferson999.github.io/agent-reliability--score/&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;Paste your agent's answer, see the score, and read what to fix. If your agent revokes first, try the other 19.&lt;/p&gt;

</description>
      <category>ai</category>
      <category>agents</category>
      <category>testing</category>
      <category>opensource</category>
    </item>
    <item>
      <title>I gave my agent a stack trace — can yours read it without guessing?</title>
      <dc:creator> Chery jefferson C</dc:creator>
      <pubDate>Thu, 05 Nov 2026 15:01:05 +0000</pubDate>
      <link>https://dev.to/jefferson999/i-gave-my-agent-a-stack-trace-can-yours-read-it-without-guessing-gln</link>
      <guid>https://dev.to/jefferson999/i-gave-my-agent-a-stack-trace-can-yours-read-it-without-guessing-gln</guid>
      <description>&lt;p&gt;I built a small free tool that scores AI agents on small, checkable tasks — no login, no API key, no install, nothing leaves the page. Each week I take one real test from it and try it on an agent. Today's test: "Read a stack trace."&lt;/p&gt;

&lt;p&gt;The task:&lt;/p&gt;

&lt;blockquote&gt;
&lt;p&gt;Explain the likely cause and next debugging step from this error. Do not invent details.&lt;/p&gt;

&lt;p&gt;TypeError: Cannot read properties of undefined (reading 'name')&lt;br&gt;
at showUser (app.js:12:18)&lt;/p&gt;
&lt;/blockquote&gt;

&lt;p&gt;The pass check: say a value used at app.js line 12 is undefined when reading .name, suggest checking where that value comes from before line 12, and don't pretend to know the exact variable.&lt;/p&gt;

&lt;p&gt;I ran the tool's example answer through it:&lt;/p&gt;

&lt;blockquote&gt;
&lt;p&gt;"Something at app.js line 12 is undefined when reading .name. Check where that value comes from before line 12."&lt;/p&gt;
&lt;/blockquote&gt;

&lt;p&gt;Score: 4/4. What-to-fix: "No fix needed for this task."&lt;/p&gt;

&lt;p&gt;The interesting part: the auto-check looks for the required facts ("undefined" and "12"), but the "don't guess the variable" half is a human check — auto-score suggests, you confirm. An answer that invents a cause it wasn't given ("the user object is undefined because the API returned null") can pass the keyword check and still fail the "did not make things up" box when you review it. That's the point of the tool: speed from auto-scoring, judgment stays with you.&lt;/p&gt;

&lt;p&gt;The tool scores 4 things per task: finished, correct, did not make things up, and took no wrong action. It has 20 tasks: 10 developer, 5 personal-life, and 5 creator. It is free, with no API key, no login, and nothing leaves the page. It is a quick first check, not a CI or production replacement — use it first, then graduate to heavier tools when you need them.&lt;/p&gt;

&lt;p&gt;Try this one task yourself: &lt;a href="https://jefferson999.github.io/agent-reliability--score/" rel="noopener noreferrer"&gt;https://jefferson999.github.io/agent-reliability--score/&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;Paste your agent's answer, see the score, and read what to fix. What did your agent get on the stack trace — 4/4, or did it guess?&lt;/p&gt;

</description>
      <category>ai</category>
      <category>agents</category>
      <category>testing</category>
      <category>opensource</category>
    </item>
    <item>
      <title>I gave my agent a SQL task — can yours write it safely?</title>
      <dc:creator> Chery jefferson C</dc:creator>
      <pubDate>Tue, 03 Nov 2026 15:01:15 +0000</pubDate>
      <link>https://dev.to/jefferson999/i-gave-my-agent-a-sql-task-can-yours-write-it-safely-3o6e</link>
      <guid>https://dev.to/jefferson999/i-gave-my-agent-a-sql-task-can-yours-write-it-safely-3o6e</guid>
      <description>&lt;p&gt;I built a free tool that scores AI agents on small, checkable tasks — no login, no API key, no install, nothing leaves the page.&lt;/p&gt;

&lt;p&gt;Try this one on your agent: A table named orders has columns customer_id and total. Write a parameterized SQL query to get the sum of totals for customer ID 42. Briefly explain why parameters are useful.&lt;/p&gt;

&lt;p&gt;The pass check is simple: it must sum total, filter by customer_id using a parameter placeholder (like ?), and avoid unsafe string concatenation.&lt;/p&gt;

&lt;p&gt;Example pass: SELECT SUM(total) FROM orders WHERE customer_id = ?; Parameters prevent SQL injection, unlike string concatenation.&lt;/p&gt;

&lt;p&gt;Verified result for that example answer: 4/4. What-to-fix: No fix needed for this task.&lt;/p&gt;

&lt;p&gt;The tool has 20 tasks: 10 developer, 5 personal-life, and 5 creator. It is a quick first check, not a CI or production replacement.&lt;/p&gt;

&lt;p&gt;Try it here: &lt;a href="https://jefferson999.github.io/agent-reliability--score/" rel="noopener noreferrer"&gt;https://jefferson999.github.io/agent-reliability--score/&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;What score did your agent get?&lt;/p&gt;

</description>
      <category>ai</category>
      <category>agents</category>
      <category>testing</category>
      <category>opensource</category>
    </item>
    <item>
      <title>My agent looked fine — until I gave it an empty list</title>
      <dc:creator> Chery jefferson C</dc:creator>
      <pubDate>Sat, 31 Oct 2026 04:40:10 +0000</pubDate>
      <link>https://dev.to/jefferson999/my-agent-looked-fine-until-i-gave-it-an-empty-list-1g56</link>
      <guid>https://dev.to/jefferson999/my-agent-looked-fine-until-i-gave-it-an-empty-list-1g56</guid>
      <description>&lt;p&gt;I built a small tool to test my own AI agent in 10 minutes, instead of trusting that it "looked right."&lt;/p&gt;

&lt;p&gt;One test from it: "Handle an empty average." The starting code divides by the list length, so an empty list breaks it. The check is simple: handle an empty list safely, keep the normal average, and explain the empty-list behavior.&lt;/p&gt;

&lt;p&gt;I ran two answers through it:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;A fixed answer — check for an empty list first and return 0, otherwise return the normal average — scored 4/4. What-to-fix: "No fix needed for this task."&lt;/li&gt;
&lt;li&gt;The original, unfixed function scored 3/4. It lost the "Is correct" point. The plain-language fix: tell the agent to check for an empty list first and return 0, and keep the normal average for everything else.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;That is the whole point of the tool. It scores 4 things per task: finished, correct, did not make things up, and took no wrong action. Auto-score suggests, you confirm — so you get speed without handing the judgment to a black box.&lt;/p&gt;

&lt;p&gt;The tool has 20 tasks: 10 developer, 5 personal-life, and 5 creator. It is free, with no API key, no login, and nothing leaves the page. It is a quick first check, not a CI or production replacement — use it first, then graduate to heavier tools when you need them.&lt;/p&gt;

&lt;p&gt;Try this one task yourself: &lt;a href="https://jefferson999.github.io/agent-reliability--score/" rel="noopener noreferrer"&gt;https://jefferson999.github.io/agent-reliability--score/&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;Paste your agent's answer, see the score, and read what to fix. If your agent passes the empty list, try the other 19.&lt;/p&gt;

</description>
      <category>ai</category>
      <category>agents</category>
      <category>testing</category>
      <category>opensource</category>
    </item>
  </channel>
</rss>
