Skip to main content

feedback-frameworks

Use when designing feedback rituals or training managers on giving and receiving feedback — covers SBI, Radical Candor, conditions for effective feedback.

跳到安装

来源信息

仓库
ccashwell/agentic-hr
最近来源活动
2026年4月25日 23:49
检测到的 SKILL.md 语言
英语
星标
0
分支
0

安装方式

默认使用会先检查来源的 Prompt;你也可以切换为直接命令,或下载本地副本。

检查来源文件

决定是否安装前,请先阅读 SKILL.md,以及 SkillsMP 当前展示的配套文件。

正在显示 SKILL.md

SKILL.md
来源说明 · 只读预览
name
feedback-frameworks
description
Use when designing feedback rituals or training managers on giving and receiving feedback — covers SBI, Radical Candor, conditions for effective feedback.
# Feedback Frameworks ## When Feedback Helps and When It Hurts Kluger & DeNisi's 1996 meta-analysis is foundational: feedback improves performance about ⅔ of the time and *hurts* performance about ⅓ of the time. The conditions for effective feedback: - **Specific behavior**, not general character - **Sincere**, from someone who cares about the outcome - **Timely**, close to the event - **Low-stakes**, separated from rating and pay - **Skilled delivery** When these are missing, feedback can backfire. The fix is not "more feedback" — it's better conditions. ## SBI (Situation-Behavior-Impact) The most usable framework for everyday feedback: - **Situation**: where, when, what context - **Behavior**: what specifically did they do - **Impact**: what was the effect Example: > "In yesterday's design review (S), when you cut Maya off three times in a row (B), I noticed she stopped contributing for the rest of the meeting (I). Was that the outcome you wanted?" The closing question is optional but powerful — it invites reflection rather than defensiveness. ## Radical Candor (Kim Scott) Two axes: - **Care personally**: do they know I have their back? - **Challenge directly**: am I willing to say what I see? The four quadrants: - **Radical Candor**: high care + high challenge - **Ruinous Empathy**: high care + low challenge (most common; nice but unhelpful) - **Manipulative Insincerity**: low care + low challenge (passive-aggressive, gossip) - **Obnoxious Aggression**: low care + high challenge (effective short-term, corrosive long-term) Most managers err toward ruinous empathy. Train for both axes. ## Specific Praise Praise is feedback too. Same rules: Bad: "Great job on the launch!" Good: "The way you sequenced the launch — handling the partner objections first, then the comms team — meant we hit the date without surprises. That sequencing was the move." The second is replicable, repeatable, motivating. The first is forgettable. ## Receiving Feedback A separate skill (Stone & Heen, *Thanks for the Feedback*). Three triggers that derail receiving: - **Truth triggers**: "this is wrong" — focus on what's right rather than processing - **Relationship triggers**: "consider the source" — dismissing based on who's giving - **Identity triggers**: "this means I'm bad" — defending self rather than considering Counter: - Distinguish appreciation, coaching, evaluation feedback - Notice the trigger; pause before responding - Ask for one specific example - "Thank you" is the only required response in the moment; processing happens after Train managers on receiving as well as giving. ## Cadence - **Real-time**: SBI in the moment when behavior is fresh — primary - **Weekly 1:1s**: standing time for feedback exchange — both directions - **Quarterly**: development conversation; broader patterns - **Cycle / formal**: confirmation of patterns; no surprises The annual review should contain zero new information. If it does, the manager has failed in the year-round. ## Upward Feedback The kind managers receive least. Engineer for it: - Direct invitation in 1:1s ("what should I start, stop, continue?") - Anonymous channels (skip-level, surveys) - Manager scorecards on cadence - Cultural permission to push back (modeled by senior leaders) Without upward feedback, managers operate on filtered information. ## Peer Feedback For development: useful, asynchronous prompts (3–5 questions tied to competencies). For evaluation: cautious. Use as input to manager assessment, not as a vote. Watch for sycophancy and retaliation. ## 360s Use with care. For development: well-suited. Aggregate input from manager, peers, reports, partners. Recipient owns the action plan. For evaluation: high risk. Easily a popularity contest. If used: - Anchored to specific competencies - n ≥ 5–8 raters - One input, not the rating - Not used alone for pay or termination decisions ## Common Failures - Vague praise / vague criticism — both information-free - Sandwich feedback (positive-negative-positive) — recipients learn to brace for the middle - Hoarding feedback for the annual - Manager-only-down feedback (no upward) - 360s used punitively - Feedback culture as code for harshness or for niceness — both miss the point ## Manager Training Module Suggested content: 1. Why feedback fails (Kluger & DeNisi) 2. SBI in practice (recorded examples) 3. Radical Candor framework 4. Specific praise practice 5. Receiving feedback 6. Role-plays with peers 7. Calibration on what "specific" looks like Allow 3 hours minimum; role-play heavy. ## Cross-References - `feedback-systems-designer` agent - `manager-enablement-coach` agent - `one-on-ones` skill - `difficult-conversations` skill ## Key References - Scott, K. (2017). *Radical Candor*. - Stone, D., & Heen, S. (2014). *Thanks for the Feedback*. - Kluger, A. N., & DeNisi, A. (1996). The effects of feedback interventions on performance. - Buckingham, M., & Goodall, A. (2019). "The Feedback Fallacy." *HBR*.
在 GitHub 查看