[kalil0321]

Fixes #140

Why

In agent mode, --url was validated and stored, but it never reached the agent. The agent only saw --prompt, so it did not know where to start unless the prompt named the site.

What

• The 4 agent engines (Claude, OpenCode, Copilot, Cursor) now accept start_url, and run_auto_capture passes it to them.
• When a URL is given, the user message gets a <start_url> block right after <mission>. The text comes from the new prompts/auto/start_url.md.
• When no URL is given, the message is unchanged.

Verification

• New tests:
  • the block is rendered for each of the 3 templates;
  • without a URL, the template output is unchanged;
  • OpenCode renders the block;
  • -u reaches the engine for all 4 SDKs.
• The new tests fail without the fix.
• pytest (1038 passed), ruff check.
• Real run with the issue's command (sdk: claude, chrome-mcp, --headless):
  • the agent's first navigation was to https://jobs.ashbyhq.com/openai;
  • the generated client returned 794 postings.

Out of scope

Reporting status: "ok" when no client is written. That is handled separately.

Confidence Score: 5/5

Safe to merge within the reviewed scope. The URL reaches the final user prompt for all four engines, and omitting it preserves existing prompt output.

Summary

Passes --url into the agent message for Claude, OpenCode, Copilot, and Cursor.

• Agent prompts now start at the supplied URL.

Acknowledged scope: kalil0321 explicitly deferred reporting success when no client is written to separate work.

No blocking issues were found; this change is safe to merge within the reviewed scope.

_{Reviews (1) · Last reviewed commit: "fix(agent): pass --url to the agent prom..." · Reviewed by Greptile}

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	🔄 Running since 2026-10-29T03:31:50.571714Z	90af85b	New commits

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[cubic-dev-ai]

All reported issues were addressed across 11 files

<sub>Tip: Get a Slack DM when cubic finishes reviewing your PRs. Connect Slack

View guided diff | Turn on auto-fix | Re-trigger cubic</sub>

[cubic-dev-ai]

1 issue found across 6 files (changes from recent commits).

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. When an issue isn't valid or won't be fixed in this PR, reply in its thread with the reason and then resolve the thread. If appropriate, use sub-agents to investigate and fix each issue separately.


<file name="src/reverse_api/prompts/auto/user_agent_browser.md">

<violation number="1" location="src/reverse_api/prompts/auto/user_agent_browser.md:62">
P2: This code-fenced command is not valid shell: Bash parses `<target-url>` as redirection and fails before `agent-browser` runs. Render the URL from `<start_url>` as a quoted argument, or use a shell-safe placeholder and explicitly tell the agent to substitute it.</violation>
</file>

<sub>Reply with feedback, questions, or to request a fix.

View guided diff | Turn on auto-fix | Re-trigger cubic</sub>

[cubic-dev-ai]

P2: This code-fenced command is not valid shell: Bash parses <target-url> as redirection and fails before agent-browser runs. Render the URL from <start_url> as a quoted argument, or use a shell-safe placeholder and explicitly tell the agent to substitute it.

Prompt for AI agents

Check if this issue is valid — if so, understand the root cause and fix it. When an issue isn't valid or won't be fixed in this PR, reply in its thread with the reason and then resolve the thread. At src/reverse_api/prompts/auto/user_agent_browser.md, line 62:

<comment>This code-fenced command is not valid shell: Bash parses `<target-url>` as redirection and fails before `agent-browser` runs. Render the URL from `<start_url>` as a quoted argument, or use a shell-safe placeholder and explicitly tell the agent to substitute it.</comment>

<file context>
@@ -59,7 +59,7 @@ Start HAR recording **before** navigation and confirm it succeeded (abort if thi
 {agent_browser_shell} network har start
 # verify exit code 0; if non-zero, diagnose with `{agent_browser_shell} doctor` and stop
-{agent_browser_shell} open https://example.com
+{agent_browser_shell} open <target-url>
 {agent_browser_shell} snapshot -i --json
 # … iterate …
</file context>