Tessl • The management layer for AI agent skills

Blog Docs Log inGet started

SkillPublic

reverser-ghidra

Analyze an authorized binary through the active Ghidra runtime and verify navigation in the mirrored GUI.

Quality

75%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./packages/decepticon/decepticon/skills/standard/reverser/ghidra/SKILL.md

Quality

Content

86%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The body is an efficient, operationally dense instruction set: specific tool APIs, exact paths, explicit verification and failure-reporting rules, and clean separation between hosted and standalone runtimes. Its only real weakness is that workflows are expressed in prose rather than as enumerated steps with worked call examples.

Dimension	Reasoning	Score
Conciseness	The body is ~20 lean lines across two sections with zero concept padding — it never explains what Ghidra or CodeBrowser are, and every sentence carries operational guidance (tool names, paths, validation rules). Matches the lean/efficient anchor; there is essentially nothing to trim.	5 / 5
Actionability	Concrete, named tool calls with signatures (`rea_open_binary`, `rea_find_tools(binary_path, query)`, `rea_call_tool(binary_path, name, arguments_json)`, `rea_status`) and exact paths (`/workspace`, `/workspace/.analysis/rea/<runId>/`) make the guidance mostly executable. Minor gaps keep it below 5: no example arguments or a worked call sequence for the hosted mode, and the hosted section describes steps more than it prescribes them.	4 / 5
Workflow Clarity	Both modes have a discernible sequence (discover/open → inspect tools via rea_find_tools → call tools → verify in GUI mirror → record hash/Evidence ID/address) plus an explicit validation checkpoint and error-recovery loop ("Inspect the GUI result before saying the operator saw the location. If following failed, keep the REA analysis result and report the lack of visual confirmation separately"). It misses 5 only because the sequence is prose-implied rather than clearly enumerated.	4 / 5
Progressive Disclosure	The skill is under 50 lines, needs no external references (none exist in references/, scripts/, or assets/), and is organized into two clearly headed sections ("Hosted engagement", "Standalone OSS runtime") with content appropriately kept inline. Per the simple-skill scoring note, this matches the top anchor.	5 / 5
	Total	18 / 20 Passed

Description

53%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description is concise, third-person, and names a distinct tool, but it under-communicates breadth (no decompile/import/cross-reference actions) and entirely lacks trigger guidance in the description field itself. The trigger terms do exist in metadata.when_to_use ("ghidra binary import functions decompile cross references codebrowser GUI"), but the description — the field evaluated and surfaced for skill selection — reads as what-only.

Suggestions

Add an explicit trigger clause to the description, e.g. "Use when the user mentions Ghidra, decompiling, disassembling, or reverse-engineering a binary."

Fold the natural trigger terms from metadata.when_to_use (decompile, disassemble, import, functions, cross references) into the description so skill selection can rely on them.

Broaden the action list beyond "analyze" and "verify navigation" — e.g. "import a binary, list functions, decompile code, and follow cross references" — to improve specificity and completeness.

Dimension	Reasoning	Score
Specificity	The description names the domain ("Ghidra runtime", "binary") and two concrete actions ("Analyze an authorized binary", "verify navigation in the mirrored GUI"), but stops short of comprehensive coverage — import, decompilation, function listing, and cross-references are never mentioned, matching the 1-2-concrete-actions anchor.	3 / 5
Completeness	The "what" is explicit (analyze an authorized binary through the active Ghidra runtime and verify GUI navigation), but the description field contains no "Use when..." clause or equivalent trigger guidance, which caps completeness at 3 per the judging guidelines.	3 / 5
Trigger Term Quality	Relevant keywords "Ghidra", "binary", and "GUI" appear, but common natural phrases users would say — "decompile", "disassemble", "reverse engineer" — are absent from the description, so coverage is partial rather than good.	3 / 5
Distinctiveness Conflict Risk	Naming "the active Ghidra runtime" carves out a clear niche with minor overlap risk against other reverse-engineering tool skills; not 5 because the generic phrase "Analyze an authorized binary" could still contend with non-Ghidra analysis skills.	4 / 5
	Total	13 / 20 Passed

Validation

93%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
metadata_version	'metadata.version' is missing	Warning

	Total	15 / 16 Passed

Repository: PurpleAILAB/Decepticon
Path: packages/decepticon/decepticon/skills/standard/reverser/ghidra/SKILL.md
Revision: 49888ee

Reviewed: 13 days ago

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.

reverser-ghidra

Table of Contents

Is this your skill?