morluto/rea

Fork 0

mirror of https://github.com/morluto/rea.git synced 2026-11-05 06:45:22 +00:00

[GH-ISSUE #6919] [Feature] Include existing static artifact routes in feature-investigation guidance #4238

New issue

Open

opened 2026-10-31 00:47:58 +00:00 by bsakaguchi · 0 comments

bsakaguchi commented

2026-10-31 00:47:58 +00:00

Owner

Originally created by @coygeek on GitHub (Oct 23, 2026).
Original GitHub issue: https://github.com/morluto/rea/issues/6919

Before submitting

I searched the open and closed issues and pull requests for a related proposal.
I understand the contribution scope must be reviewed before integration; open questions are welcome here.
This request does not include credentials, private samples, or sensitive target data.

Area

MCP or agent integration

Current limitation

The investigate_feature prompt offers JavaScript analysis and native open_binary as its initial routes, followed by active-document and procedure operations. REA's installed skill already describes other supported starting points: inspect_android_package for APK code, inspect_managed_artifact for managed assemblies, and open_bundle_resources for Apple resources without an executable or deep provider.

A caller using the guided prompt alone does not receive those alternatives. This is source-derived evidence of narrower guidance, not an observed claim that an agent chose the wrong tool or that those analysis operations fail.

Use cases

Investigate a declared class or method in an APK using the existing Android tools.
Understand a method in a managed assembly without first choosing a native decompiler.
Inspect an Apple application's plist or Interface Builder resources when the question does not require an executable.

Desired outcome and success criteria

Expand the existing prompt's optional starting points to distinguish these supported artifact-and-question routes before suggesting native analysis.

Guidance distinguishes APK code inspection from archive-member inventory, managed metadata from native code, and Apple resource inspection from executable analysis.
Target-free operations receive their explicit path directly; active-target workflows retain their existing session requirements.
Native analysis remains an option when the question requires native behavior. The prompt does not assume a deep provider is necessary for every file or app bundle.
Route recommendations preserve host/provider limitations and refer to the connected server's advertised capabilities. A filename suffix alone does not establish an artifact's actual format or analysis result.
Rendering remains guidance-only: it neither reads the target nor starts a provider, and it preserves caller context, optional steps, Evidence discipline, and explicit unknowns.

Expected interaction with targets

No target interaction (documentation or usability only)

Target and scope details (optional)

Only the existing feature-investigation prompt and its supporting guidance/checks are in scope. The later analysis operations have their own effects and prerequisites; rendering this prompt does not perform or authorize them. No new detector, provider, automatic execution, or mandatory sequence is requested.

Verification available so far

I can suggest a sample or manual check

Verification details (optional)

Source review only; no runtime or agent-selection experiment was performed. Proposed checks: obtain the native MCP prompt for questions about APK methods, a managed assembly method, and an Apple resource-only bundle. Verify that the corresponding existing routes and their prerequisites are explained. Keep JavaScript and native-function cases as controls, and verify that prompt rendering acquires no target or provider process. Test format uncertainty explicitly instead of treating an extension as proof.

Alternatives considered

Read the installed skill alongside the prompt. It already supplies the missing routes, but a prompt-only caller has to discover and reconcile that second surface.
Create separate prompts for every artifact family. That duplicates general Evidence discipline and leaves the existing broad feature prompt narrower than the supported workflows.
Add another artifact detector. REA already has inspection operations; this request improves how existing operations are explained.

Possible direction

Bring the prompt's conditional starting points into agreement with the installed skill's artifact-and-question routing. Keep the text selective and optional rather than embedding the complete tool catalog.

References

Reviewed source: REA package 6.7.0, main revision bcd18d289845051b0f0513e11bc1aefd3f766902. The feature prompt's initial routes and the skill's existing artifact routes show the difference. The renderer defines guidance-only behavior.

Related work: #102 introduced guided prompts and PR #733 clarified JavaScript directory routing. This request extends route coverage to other already-supported static artifact workflows.

Triage assessment

Impact level: not-applicable
Assessment status: not-applicable
Impact basis: This expands optional guidance coverage. No incorrect runtime result or breach of an existing prompt contract was demonstrated.
Workaround status: not-applicable
Workaround basis: This is a feature request. Existing skill guidance and direct analysis operations remain usable today.

Originally created by @coygeek on GitHub (Oct 23, 2026). Original GitHub issue: https://github.com/morluto/rea/issues/6919 ## Before submitting - [x] I searched the open and closed issues and pull requests for a related proposal. - [x] I understand the [contribution scope](https://github.com/morluto/rea/blob/main/CONTRIBUTING.md#contribution-scope) must be reviewed before integration; open questions are welcome here. - [x] This request does not include credentials, private samples, or sensitive target data. ## Area MCP or agent integration ## Current limitation The `investigate_feature` prompt offers JavaScript analysis and native `open_binary` as its initial routes, followed by active-document and procedure operations. REA's installed skill already describes other supported starting points: `inspect_android_package` for APK code, `inspect_managed_artifact` for managed assemblies, and `open_bundle_resources` for Apple resources without an executable or deep provider. A caller using the guided prompt alone does not receive those alternatives. This is source-derived evidence of narrower guidance, not an observed claim that an agent chose the wrong tool or that those analysis operations fail. ## Use cases - Investigate a declared class or method in an APK using the existing Android tools. - Understand a method in a managed assembly without first choosing a native decompiler. - Inspect an Apple application's plist or Interface Builder resources when the question does not require an executable. ## Desired outcome and success criteria Expand the existing prompt's optional starting points to distinguish these supported artifact-and-question routes before suggesting native analysis. - Guidance distinguishes APK code inspection from archive-member inventory, managed metadata from native code, and Apple resource inspection from executable analysis. - Target-free operations receive their explicit path directly; active-target workflows retain their existing session requirements. - Native analysis remains an option when the question requires native behavior. The prompt does not assume a deep provider is necessary for every file or app bundle. - Route recommendations preserve host/provider limitations and refer to the connected server's advertised capabilities. A filename suffix alone does not establish an artifact's actual format or analysis result. - Rendering remains guidance-only: it neither reads the target nor starts a provider, and it preserves caller context, optional steps, Evidence discipline, and explicit unknowns. ## Expected interaction with targets No target interaction (documentation or usability only) ## Target and scope details (optional) Only the existing feature-investigation prompt and its supporting guidance/checks are in scope. The later analysis operations have their own effects and prerequisites; rendering this prompt does not perform or authorize them. No new detector, provider, automatic execution, or mandatory sequence is requested. ## Verification available so far I can suggest a sample or manual check ## Verification details (optional) Source review only; no runtime or agent-selection experiment was performed. Proposed checks: obtain the native MCP prompt for questions about APK methods, a managed assembly method, and an Apple resource-only bundle. Verify that the corresponding existing routes and their prerequisites are explained. Keep JavaScript and native-function cases as controls, and verify that prompt rendering acquires no target or provider process. Test format uncertainty explicitly instead of treating an extension as proof. ## Alternatives considered - Read the installed skill alongside the prompt. It already supplies the missing routes, but a prompt-only caller has to discover and reconcile that second surface. - Create separate prompts for every artifact family. That duplicates general Evidence discipline and leaves the existing broad feature prompt narrower than the supported workflows. - Add another artifact detector. REA already has inspection operations; this request improves how existing operations are explained. ## Possible direction Bring the prompt's conditional starting points into agreement with the installed skill's artifact-and-question routing. Keep the text selective and optional rather than embedding the complete tool catalog. ## References Reviewed source: REA package 6.7.0, main revision `bcd18d289845051b0f0513e11bc1aefd3f766902`. The [feature prompt's initial routes](https://github.com/morluto/rea/blob/bcd18d289845051b0f0513e11bc1aefd3f766902/src/contracts/promptContracts.ts#L95-L152) and the [skill's existing artifact routes](https://github.com/morluto/rea/blob/bcd18d289845051b0f0513e11bc1aefd3f766902/.agents/skills/reverse-engineer-anything/SKILL.md#L93-L123) show the difference. The [renderer](https://github.com/morluto/rea/blob/bcd18d289845051b0f0513e11bc1aefd3f766902/src/contracts/promptContracts.ts#L52-L91) defines guidance-only behavior. Related work: [#102](https://github.com/morluto/rea/issues/102) introduced guided prompts and [PR #733](https://github.com/morluto/rea/pull/733) clarified JavaScript directory routing. This request extends route coverage to other already-supported static artifact workflows. ## Triage assessment - Impact level: not-applicable - Assessment status: not-applicable - Impact basis: This expands optional guidance coverage. No incorrect runtime result or breach of an existing prompt contract was demonstrated. - Workaround status: not-applicable - Workaround basis: This is a feature request. Existing skill guidance and direct analysis operations remain usable today.

bsakaguchi referenced this issue

2026-10-31 01:12:11 +00:00

[PR #4238] [MERGED] fix(lifecycle): preserve ownership, cancellation, and cleanup outcomes #6141

No milestone

No project

No assignees

1 participant

Notifications

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference

morluto/rea#4238

No description provided.

Rows
Columns