CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

dev-mem

Runtime memory access and manipulation on embedded Linux via /dev/mem, /dev/kmem, and MTD devices. Covers physical memory reads/writes, kernel symbol resolution via /proc/kallsyms, live firmware patching with mtd_debug and flashcp, and bypassing memory-access restrictions (CONFIG_STRICT_DEVMEM, kernel.perf_event_paranoid).

Quality

77%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Critical

Do not install without reviewing

Fix and improve this skill with Tessl

tessl review fix ./packages/decepticon/decepticon/skills/standard/iot/dev-mem/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

71%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.

A highly actionable, command-dense skill whose destructive MTD-write workflow is properly validated with backup and verify steps. It is weakened by a broken reference to a nonexistent references/devmem-kernel-exploit.md and by deep-dive content (U-Boot env manipulation) inlined in SKILL.md instead of split into a bundle file.

Suggestions

Create references/devmem-kernel-exploit.md (or remove the link) — the References section currently points to a file that does not exist, breaking navigation and progressive disclosure.

Move the Phase 3d U-Boot env format/CRC deep-dive into a references/uboot-env.md file, keeping a short pointer and the one-line decode command in SKILL.md.

Add error-recovery guidance after the Phase 3c verify step (what to do on MISMATCH, e.g. re-erase and re-flash from the backup) to complete the feedback loop on the destructive operation.

DimensionReasoningScore

Conciseness

The body is command-dense with terse inline comments, but includes minor over-explanation of concepts Claude already knows ("MTD (Memory Technology Devices) exposes raw NOR/NAND flash", "Available on older kernels (pre-3.7)") and repeats the kallsyms grep pattern in both Phase 2 and Phase 5. Not 5: a few tokens are spent on known concepts and duplication; not 3: the padding is minor and localized, not "some unnecessary explanation" throughout.

4 / 5

Actionability

Nearly everything is copy-paste-ready: devmem2/dd reads, mtd_debug/flash_erase/flashcp write sequences, and a complete executable Python script for U-Boot env CRC patching. Minor gaps remain, e.g. the undefined placeholder "skip=$((heap_start))" in Phase 1c and the hardcoded "HEAP_START=0x00400000" in Phase 4. Not 5: a few placeholders prevent direct execution; not 3: the guidance is overwhelmingly concrete and runnable.

4 / 5

Workflow Clarity

Phases 1–5 give a clear sequence, and the destructive MTD write in Phase 3c includes explicit checkpoints: backup, sha256sum, flash_erase, flashcp write, verify-read, and "diff ... && echo WRITE OK || echo MISMATCH" confirmation. Not 5: the read-oriented phases (1, 4, 5) lack validation checkpoints and there is no error-recovery guidance if the Phase 3c verify reports a mismatch (e.g. restore-from-backup instructions).

4 / 5

Progressive Disclosure

Section headers are clear and navigable, but the sole bundle reference "references/devmem-kernel-exploit.md" points to a file that does not exist (no references/ directory), and the full U-Boot env format/CRC deep-dive in Phase 3d (~40 lines including a Python script) is inlined reference-grade content. Not 4: a broken reference and inlined deep-dive content; not 2: the section structure is genuinely good and the body is not an unstructured wall of text.

3 / 5

Total

15

/

20

Passed

Description

75%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.

A highly specific, distinctive description that names concrete tools and kernel configurations comprehensively. Its main weakness is the absence of an explicit "Use when..." trigger clause in the description field itself — trigger guidance is relegated to metadata.when_to_use — which caps completeness.

Suggestions

Append an explicit trigger clause to the description, e.g. "Use when the task involves /dev/mem, /dev/kmem, MTD writes, devmem2, kernel memory access, or live firmware patching on embedded Linux."

Fold the strongest natural trigger terms from metadata.when_to_use (devmem2, embedded memory access, live firmware patch) into the description so trigger matching does not depend on a separate metadata field.

Add user-natural phrasings like "dump physical memory" or "patch firmware in place" to broaden keyword coverage toward anchor 5.

DimensionReasoningScore

Specificity

"physical memory reads/writes, kernel symbol resolution via /proc/kallsyms, live firmware patching with mtd_debug and flashcp, and bypassing memory-access restrictions (CONFIG_STRICT_DEVMEM, kernel.perf_event_paranoid)" lists multiple specific concrete actions naming exact tools and kernel configs, comprehensively covering the skill's capabilities. Not 4: coverage is comprehensive rather than having minor gaps.

5 / 5

Completeness

The "what" is clear and detailed ("Runtime memory access and manipulation on embedded Linux via /dev/mem, /dev/kmem, and MTD devices"), but the description field contains no "Use when..." clause or equivalent trigger guidance — triggers live only in the separate metadata.when_to_use field, capping completeness at 3 per the rubric. Not 4: "when" is absent from the description field itself, not merely weakly implied.

3 / 5

Trigger Term Quality

Terms like "/dev/mem", "/dev/kmem", "MTD", "/proc/kallsyms", "mtd_debug", and "flashcp" are the natural phrases a user would say for this task, but the description field omits common variations such as "devmem2", "memory dump", or "process memory". Not 5: synonym coverage is incomplete; not 3: keyword coverage is good and domain-natural, not merely "some relevant".

4 / 5

Distinctiveness Conflict Risk

The description is anchored on uniquely niche device paths and tools (/dev/mem, /dev/kmem, MTD, mtd_debug, flashcp) that no other skill would claim, giving a clear niche with minimal conflict risk. Not 4: the trigger surface is distinct enough that even minor overlap risk is negligible.

5 / 5

Total

17

/

20

Passed

Validation

87%

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

Validation — 14 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

metadata_version

'metadata.version' is missing

Warning

referenced_paths_exist

Referenced path issues: 1 missing

Warning

Total

14

/

16

Passed

Repository

PurpleAILAB/Decepticon

Path

packages/decepticon/decepticon/skills/standard/iot/dev-mem/SKILL.md

Reviewed

Table of Contents

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.